Omniscio documentation library
Every Omniscio feature and subsystem, one page each — written for a person who has never opened the app, and for an AI with no access to this repository.
Getting Started 13 Installing Omniscio, signing in, and finding your way around for the first time.
Install & Sign In Machine Migration (moving Omniscio to a new computer) How to move a whole Omniscio install — sessions, conversations, settings, attachments and automations — to a new computer. The recommended path is the encrypted full-state snapshot you restore on the new machine in about a minute; the page also covers the manual export, choosing replace versus merge, and what to do when it goes wrong. Mandatory Claude Extensions (now all optional, and no longer installed at setup) How Omniscio used to force-install the extensions it depended on during first-run setup, and why nothing is force-installed any more. It covers what Setup v2 installs instead (CLI tools only), the launch-time banner that survives but stays quiet, and the one-time offer existing users get to drop the old plugin in favour of the built-in pipeline. Portable startup splash The two invisible phases before the real Omniscio window appears — the portable .exe's extraction phase and the app's own boot phase — and how the startup splash makes both visible: the static extraction image, the boot splash with its logo, spinner, phase checklist and live status line, and the kill switches that suppress it. Signing in to Omniscio (and having one account, not two) The sign-in screen you meet before Omniscio opens, plus the matching card at Settings → Accounts. It is separate from the Claude/Anthropic login that runs your sessions — this one is your Omniscio account, and it is what your plan, your team, and cloud sync hang off. Windows Installer (real progress + zero-click updates) Two layered Windows install artifacts: a branded bootstrap stub that is the primary download and shows real byte progress and a measured time left while it downloads, verifies and installs, and the full NSIS installer kept for offline installs and as the engine auto-updates drive. Covers the fresh-install wizard, the zero-click update apply, and uninstall — including what uninstalling does on macOS and Linux, where there is no uninstaller and nothing prompts you about the data left behind. First Run "Reply to start" — waking your first agents A calm one-time banner shown just after setup, telling you that your prepared agents are ready and waiting and that replying is the single action that starts one. It appears only when an agent is genuinely waiting, disappears on its own once you have replied to the last one, and replaces the old completion message that wrongly claimed they were already working. First Mission (guided onboarding experience) The short coached experience that runs right after a new user finishes setup: a read-only health check across performance, security, privacy, drive health and hardware, then the worthwhile fixes one at a time, each explained and approved by you. First-time setup The short sequence that takes a freshly installed Omniscio from first launch to a working session against your own code — the two-phase wizard, what each step asks for, and what you land in when it finishes. Off-machine backup prompt (a backup that survives losing the computer) A default install's only backup lives on the same disk as the database, so losing the machine loses everything. This is the gentle, consent-first prompt that invites a new user to set up a backup that lives somewhere else — what it says, when it appears, and why it never turns a backup on for you. Onboarding Guardian (a quiet hint when setup stalls) An invisible assistant that watches a new user move through setup and the interactive tour, staying silent almost all of the time and speaking up only when someone is genuinely stuck on a step. It exists so a first run that stalls produces one gentle hint instead of a dead end. Setup Assistant card (post-onboarding guided setup) The one-time inbox card a brand-new user gets after onboarding, offering a guided setup assistant: what it offers, what launching it opens, and why it is off by default until it ships. Tour & Orientation App Tour Omniscio's guided-onboarding system — the full-screen overlays and spotlight tooltips that walk you through a feature step by step, the tours on offer and where each one surfaces, the JSON that defines them, and how to author a new tour or anchor. Help & Docs Panel (in-app docs site) A built-in help browser that loads the public Omniscio documentation inside the app, with back, forward, home and an open-in-browser escape hatch — so you can read the docs beside your sessions instead of switching to Chrome. Sessions & Agents 115 Running AI sessions and agents — starting them, steering them, and orchestrating many.
Sessions Acting on a whole section at once — the “…” on a section header The small “…” button on the header of each group of sessions — Needs you, Live, Interrupted, Paused, Pinned, Scheduled, Snoozed, Saved, Archived and your own tag groups. It opens the same menu you get by right-clicking a session, pointed at everything in that one group, so you can snooze, pause, restart or archive a whole section at once without ticking rows one by one. The menu only ever offers what actually applies to the sessions in that group. Agent input modal (a session asks you to fill a value it must never see) A session that is blocked on a value only you can give — an API key, most often — can make a labeled modal appear over the app; you type it, it saves through the normal settings path, and the AI never sees what you typed. Agent Message Display (structured agent prose inline) How a session feed draws an agent's reply — substantive prose between tool calls is promoted inline while short narration folds into an activity pill, plus the settings, the copying controls, and the multi-gate fold that change what you see. Agent reply formatting and showing images The formatting an agent can put in a reply — callouts, centred text, sized images, collapsible sections, keyboard key caps — and how an agent shows you any image file on your machine, including what happens when a picture cannot be displayed. Agent self-archive When an agent finishes its work with nothing left for you to do, it can archive its own session — the row drops into the Archived section on its own, fully reversible, and stamped with an Agent (self-archived) chip. Two settings decide whether it may. Agent Status Board A live board showing, for every Claude Code session Omniscio launched, where that agent is working, what state it is in, its dev-pipeline phase and gate, and its to-do list — with one click to jump into the session. It is opt-in and view-only, a card updates when its agent does something, and the board comes straight back after a restart. App messages in the transcript (readable system cards) Messages the app itself sends into an agent's session (an Inbox decision, an Auto-lander hand-back, a Gate watch or Land watch result, a scheduled wake-up) are written for the agent. They now render as short cards that say what happened in plain words, with the original one click away and the agent's copy untouched. Archive a session Archiving takes a session out of the sidebar's main list and parks it in the Archived sub-section without deleting anything — how to archive from the keyboard, the menus, middle-click or mobile, how undo works, and what happens under the hood. Asides (side question without polluting main thread) Side questions you can ask an agent without polluting the main thread: how to enter aside mode and leave it, what the forked one-shot spawn behind it costs, where its settings live, and the limits — including why an agent can never send one. Attach a file How to put images, PDFs, text and Office documents onto a message — the paperclip, drag-and-drop, paste and keyboard shortcut — which file kinds are accepted or refused, the size caps, and what happens to each kind after Send. Auto-open a session that needs you (idle sessions panel) Auto-open a session that needs you — the embedded Sessions panels open a waiting session's chat by themselves when you arrive on an empty pane, the rules that stop it interrupting you, the order it picks between several, and why it never applies on your phone. Branch header The optional git branch badge in a session header: showing the current branch next to the session name, the worktree icon and tooltip, the global and per-project toggles that hide it, and how the branch value is watched and refreshed. Browse archived sessions How to open and move between a project's archived sessions: the Archived sub-section, the read-only archived view, swipe and J / K navigation, where an opened session lands, and the separate navigation list that makes prev and next work. Bulk Select Sidebar Items Selecting several rows at once in the sidebar — sessions, or entries in the unified inbox — and then acting on all of them with one keystroke or one right-click. Covers the Shift+J/K and click-modifier gestures that extend a selection, what each bulk action does, and which single-row actions gray out when more than one row is picked. Catch-Up Card A pinned four-line summary that appears above a session's transcript while that session is waiting on you, so you can pick the conversation back up without rereading it. Covers what each line is, when the card appears and disappears, how it collapses, what each regeneration costs, and the fact that it is currently switched off for everyone. Chat attachments (PDFs, text docs, images, Office docs, pasted text) How files and pasted text reach Claude from the Omniscio chat composer: the paperclip, drag and drop, paste and screenshot routes, the five buckets an attachment is sorted into, the chips you see, the size caps, the large-draft warnings, and the send paths that carry attachments. Chat attachments (PDFs, text docs, images, Office docs, pasted text) (part 2) Part 2 of the Chat attachments page: the send pipeline that partitions every attachment and delivers it, how non-Claude engines receive attachments as files in the workdir, how a large paste becomes a chip instead of flooding the composer, and where the attachment files live on disk. Clear a hub's waiting items — middle-click a hub Clearing everything a sidebar hub is waiting on you for, with a middle-click on that hub's row — or on the Inbox row to do every hub at once. What the confirmation names, which rows count as waiting on you, the few things that cannot be undone, and how the sessions come back with one Undo. Code blocks that are really prose Why a fenced block that holds prose — a prompt to copy, a quoted message, a document excerpt — used to read as garbage, and how Omniscio now reflows it into flowing paragraphs while leaving real code, lists, and tables exactly as the author wrote them. Compaction Summary (expandable inline) What you see when a session's context window fills and Claude writes a condensed summary of the conversation so far — the "Conversation compacted" divider, and the Show summary link that expands the full text inline. Covers how to open and copy it, when the link is not offered, and how history loads after a compaction. Context Details (see what's filling the agent's context window) A popover that breaks down what is filling a session's context window right now — the summary row, the expandable per-source sections, and the refresh button — so you can decide whether to keep going, clear, or start fresh. Covers the two ways to open it, the optional donut-ring indicator, why it can be empty, and where the numbers come from. Context Usage Warnings The two optional in-chat notes that appear when a session's context window is filling up — one at 75% as a heads-up, one at 90% as a warning that the conversation is about to be compacted. Covers how to switch them on, where the on/off controls sit in the chat itself, and why a warning never repeats until a real compaction re-arms it. Copy / export a session as Markdown Taking a session's conversation out of Omniscio — copying it to the clipboard, saving it as Markdown, Word, PDF or a Google Doc, sharing it as a web link, what "visible-only" leaves out, and how the size estimate and the since-last-compaction slice work. Detached session window (Pop Out) Pop a session out of the main Omniscio window into one of its own, so you can keep watching or driving it beside other work — how to open one, what it shares with the session it came from, and how to put it back. Edit a user message You cannot change a message you already sent inside its own session, but you can fork from that point into a new one with your edit applied — what the app offers instead, and the workarounds when you genuinely need the original to change. Hide background check-ins Collapses the trailing status message a session posts when its own background job wakes it back up — a build, a test run, a Monitor watch, a scheduled wake-up — into a single quiet one-line marker you can click to expand. Covers which turns fold, which never do, the setting that controls it, and the rules that decide. Image Bridge For Text-Only Models An optional session setting that lets a model which cannot read images still work from one: a vision-capable model writes a brief description first, and that brief travels to your chosen text-only model alongside your message. Interrupted sessions section (sidebar) A dedicated sidebar section, above Live Sessions, holding the sessions whose process stopped on its own rather than by your hand — a crash, a finish, or an unexpected exit. The header counts them and each row carries a red dot, because Interrupted means something ended that you did not end. Link hover web-title preview (hover a chat link → see the page title) Hovering a web link in a session or chat shows a tooltip with that page's real title — the article headline, the repo name, the video title — instead of the bare URL, so you know what you are about to open before you click. The title is fetched once and remembered; while it loads the tooltip shows the domain. Mermaid diagrams (draw a diagram inside a chat message) A mermaid code block in a chat message is drawn as a real diagram — flowchart, sequence, state or class — instead of showing you the raw text. Theme-matched, scrollable, always on, with a view-source toggle, and rendered as untrusted content so a diagram can never inject anything into the app. Message image carousel (a run of images becomes one swipeable viewer) When an AI reply shows three or more images in a row, Omniscio collapses them into a single swipeable carousel with a corner count instead of a tall stack you scroll through. Covers how you page through them on a phone and on desktop, and which runs of images deliberately stay inline. More replies below (a count of what you have not read yet) When you reopen a busy session Omniscio lands you on the most important unread reply, and this little indicator tells you how many real replies are still waiting below it. Tapping it walks you forward one reply at a time, so you never skip past the buried answer you were looking for. Moving a session between projects Re-file a session under a different project — one at a time, or several at once — from its own menu or a single keystroke, live on every window and on your phone at the same time. Covers the deliberate exception for a session that has already started, and the Undo that puts everything back exactly as it was. Native slash commands (run /model, /clear, /cost for real) Type Claude Code's own slash commands — /model, /clear, /cost, /context, /status — in the Omniscio composer and have them run in the underlying Claude Code session rather than being sent to the model as text. What each command does, which sessions it works in, and the setting that turns it off. Omniscio awareness prompt — sessions know they're in Omniscio An on-by-default note Omniscio adds to every session's system prompt so the agent knows it is running inside Omniscio, which everyday requests are really Omniscio actions, and how to reach Omniscio's own controls to do them — what the note tells the agent, when it is skipped, and how to turn it off. Opening a media link (it opens in your own player) When an agent's message links to a binary media file — audio, video, an archive, a PDF or an HTML file — clicking the link hands it to your operating system's default app for that kind of file instead of trying to draw it inside the app. Covers which extensions route this way and what deliberately does not. Pasting rich text (formatting survives the trip into chat) Paste from Google Docs, Word on the web, a web page or Notion and Omniscio converts the formatting to Markdown, so headings, bullets, bold and italic, code, quotes and tables arrive intact in the composer instead of as a wall of characters. Includes how to force plain text and what is deliberately dropped. Pause or stop a session (four ways, and which one you want) Omniscio has four different ways to stop a session from running, and they do genuinely different things — one freezes it so it can pick up exactly where it left off, one ends the turn, one kills it for good. This page is the one that tells you which to reach for, and what each does to the work in flight. Peek viewer (look at a file without leaving what you were doing) A slide-over panel that shows a file in place — text, code, images — so you can read what an agent is talking about without opening an editor or losing your place. Covers opening one from a link or a path, finding text inside it, how it works on a phone, and how a path is resolved to the right copy of a file. Per-model thinking level (a default reasoning level per model) Give each model its own default reasoning level, so every new session inherits the right amount of thinking from the model it uses — heavy for a model that benefits from it, light for one that is meant to be fast. A supplement to your single Thinking Level, never a replacement, and anything you choose for a specific session still wins. Plain Speak (a plain-English summary beside every reply) A short plain-English rewrite delivered alongside an agent's reply, so you can tell what happened without reading the technical text underneath it. This page covers what it is and how the card reaches you; the parts below cover what you see, the controls, and the internals. Plain Speak — gates, cost and privacy (part 3) Part 3 of the Plain Speak page: the reports a build pipeline gets, exactly what the rewrite is allowed to see, the built-in prompt it is locked to, the master switch and the spending controls behind it, the log of decisions it made, and what leaves your machine. Plain Speak — implementation notes (part 4) Part 4 of the Plain Speak page: the implementation notes for anyone working on the code — where the rewrite is computed, the shape of the output, how a failure is handled, and the tests that pin the behaviour. Plain Speak — what you see (part 2) Part 2 of the Plain Speak page: the rewrite as it appears on a message, what happens when a reply is flagged as unsafe to rewrite, and the whole experience of waiting — what is on screen while the rewrite is still being computed. Question Widget (inline answer pills with stable position) Omniscio's inline Question Widget: what happens when an agent asks a multiple-choice or yes/no question, where its settings live, how to answer with pills, keys or a typed reply, and how the layout, the option filter and the saved state behave while you work through it. Question Widget (inline answer pills with stable position) (part 2) The second half of the Question Widget page: the Plain Speak canonical-priority trigger and its aggressive numbered-bullet mode, the format hint Omniscio adds to the first message of a session, the v3 parser's history, how widget state is saved, and the file-level pointers an engineer touching this code needs. Queue a message ("send after the agent finishes") How to line up follow-up messages that send themselves the moment the agent finishes its current step. Covers where the option lives in the composer, how the queued list is opened and managed, and the limits worth knowing before you rely on it — the 24-hour wait, the refusal on an ended session, and how several messages go over together in one turn. Quiet chip (the "16m" clock badge on a session row) A small grey clock badge, reading 16m or 45m or 2h, at the right-hand end of a session row. It means exactly one thing: the session is running but has produced no output for that long. It is a silence timer, not an error and not a progress bar, and it clears the instant the session speaks again. QW miss reporting (Report missed question widget) How to flag an agent message that should have rendered as a clickable question widget but did not — where the menu item lives, what the report bundles and ships to the parser team, and the privacy boundary around what leaves your machine. Re-run from here (edit a message & re-run in a new session) Pick any message in a conversation and re-run the chat from that point in a brand-new session: reword one of your own messages and resend it, or write a fresh next message from an AI reply. Omniscio forks a session that shows the whole conversation up to the point you picked, and never changes the original. Real Conversation Layout (hide the tool noise, keep your first message pinned) The chat panel shows just the real conversation: your messages and the agent's actual replies, with each turn's tool work folded into a collapsed activity toggle. This page covers what you see, the auto-continues that are hidden, when the layout applies and when it does not, and the setting and kill switch behind it. Real Conversation Layout — part 2 (how the panel decides what to render) The decision layer behind the Real Conversation Layout: which rows the chat panel is allowed to show, how it decides a turn was real conversation rather than plumbing, the opt-in marker that pins a final answer, and the gate reports and no-answer turns it handles specially. The visible result is described in part 1. Real Conversation Layout — part 3 (the compact clamp and the code map) The last details of the Real Conversation Layout: how the cold-mount fetch interacts with the compact clamp, and a map of the code the whole feature lives in. Part 1 covers the visible layout and its setting; part 2 covers how the panel decides what to render. Rename a session Every session has a title you can change to anything up to 100 characters, from three places: right-clicking the title at the top of an open chat, the three-dot menu in the session header, or right-clicking the row in the sidebar. Renaming only changes the display name, it stops the automatic titler overwriting your own name, and it is undoable. Reporting a bad Plain Speak rewrite A "Report bad overlay" item in an agent message's menu, shown while that message is displaying its plain-English rewrite: one short note and the app attaches the original message and the rewrite for you. It is how a bad rewrite gets back to the people tuning the pipeline, with the message text sent and nothing else. Run a file from chat (right-click a runnable path → Run / Run as administrator) When a chat message mentions a runnable file by its path, Omniscio turns that path into something you can run without copying it into a terminal yourself: right-click the path and pick Run or Run as administrator, confirm the exact path, and Omniscio launches it in a terminal window that stays open. Save a session Saving a session bookmarks it into a collapsible Saved section in the sidebar, between Snoozed and Archived. The session keeps its current status — saving is an organizational flag like starring an email, not a status change — and a saved session that needs you surfaces in Needs You instead. Scroll position memory When you scroll up in a session and then leave it, Omniscio remembers where you were and returns you to the same scroll position when you come back — silently, with no toast or animation. The remembered position lives in memory only and is forgotten on app restart. Send a message The composer is the textarea pinned to the bottom of an open session. Type a message and press Enter, and Omniscio delivers it to that session's running Claude process as one operator turn, with the reply streaming into the chat above. Attachments, AI-suggested replies, snippets, the skill menu and Send Later all live here too. Send Later (scheduled responses) Send Later queues a reply you write now for delivery at a time you pick, carrying whatever images, PDFs and pasted text were in the composer when you scheduled it. It uses the same time picker as Snooze, is reachable from the local control server, and raises an inbox alert if delivery finally fails. Sender Run Cards (back-to-back messages become one quiet line that opens to a swipeable deck) When two or more messages arrive back to back, the conversation shows one quiet line naming how many and who from, which opens to a swipeable deck of cards. Every card holds the real message, a counter tells you where you are, and Show all spreads the burst into a plain list. On by default. Session Files & Links (what this session made) A read-only, per-session list of every file the agent created or edited in that conversation, and every link it handed you — newest first, one clickable row each. It is turned on from Labs and stays off by default; once on, a Files button in the session header opens the panel, and the same list is deliberately available on your phone, where there is no file tree at all. Session handoff (carry a long chat into a fresh session) Session handoff turns one long, cluttered session into a fresh one that is already caught up: same project, folder and branch, carrying a summary that keeps the conclusions and lists the dead ends as things not to retry. One click in a session's overflow menu does it, and the old session is left completely untouched. Session hibernation (a heartbeat session goes quiet between beats) Sessions that wake on a schedule just to find nothing to report would otherwise land in your inbox every single time. Hibernation lets such an agent stay green and running quietly until its next beat instead of finishing into your inbox. It is on by default, requested by the agent, and saves your attention, not money. Session isolation How to give a session its own private copy of your project instead of letting it edit your real folder — where the switches live, what that copy contains and what it leaves behind, how the work comes back when the session ends, and what happens when a copy cannot be made. Session provenance — trace inbox actions and spawned sessions to their origin When many agents run at once it is not obvious which one caused an inbox item, a spawned session or an injected turn. Provenance answers that with a clickable link back to the origin: a generated-by line on approvals and alerts, a spawned-by note atop a new session, a children chip on the parent, a via-CLI tag on injected turns. Session Refill Governor — "keep this many sessions running" When fewer sessions are working than you want, this quietly restarts ones that were interrupted, meaning work that ended abnormally and left something unfinished, until you are back at your target. It checks every five minutes, is off by default, and never touches a turn that finished cleanly. Reopening sessions you closed yourself is a separate switch. Show system messages (per-session diagnostic toggle) The chat panel's default view — the real-conversation layout — hides the "plumbing" rows that Omniscio generates around your actual conversation, so a long session reads as the dialogue you actually had. Most of the time that's what you want. But sometimes you need to confirm the plumbing is working: did the OAuth retry actually fire? did the account-switch recovery happen? is the away-mode auto-reply going out? Show system messages is a per-session toggle that reveals those hidden rows for o... Snooze a session Snooze temporarily hides a session from the sidebar and attention lists until a time you pick — useful for "I'll deal with this later." When that time arrives (or you unsnooze early), the session returns to your inbox as "needs you" so you actually see it again — even if it was paused, ended, or archived when you snoozed it. It is different from pause: snooze only hides the session in the UI, while pause actually freezes (kills) the underlying Claude CLI process. Snooze a session (part 2) The snoozedUntil field on Session (an ISO timestamp or null) is the single source of truth for the snoozed/active flip. The chat marker is implemented as a system message in the conversationmessages table with metadata.kind === 'snooze-marker'. Its metadata shape is the SnoozeMarkerMetadata interface: snoozedUntil — ISO timestamp the snooze ends. Immutable after insert. note? — optional plain-text note, ≤500 chars, newlines preserved. Immutable after insert. Start a new session A session is one chat with Claude Code running inside a specific project's folder — the agent can read, edit, and run code against that folder, and your back-and-forth lives in the session's transcript. You can have many sessions open across many projects at the same time. Starting a new session means: pick a project, hit the launch button (or Ctrl+T / N), and Omniscio spawns a fresh claude CLI process inside that project's folder using your active Claude account. Stop, restart, pause, archive, message, or refill many sessions at once The Manage sessions modal: one picker with a view per action — Stop, Restart, Pause, Archive, Message, Refill and Queued — so you can act on a hand-picked set of sessions instead of all or nothing. Covers where to open it, what each view lists, the filter and selection controls, what is undoable, and the safer narrow rules Stop and Restart follow. Stop, restart, pause, archive, message, or refill many sessions at once (part 2) How the bulk Pause, Stop, Restart, Archive and Broadcast actions are actually wired: which store call each button loops over, the selectors that decide which sessions each view offers, the single terminate-many request Stop sends, the delivery channel a broadcast reuses, and the contracts that lock the whole rule set. Tag a session Session tags are short colored labels you attach to a session for quick categorisation — things like urgent, bug or client-acme. The free-form flavor is what this page covers: short strings you type into the picker on the spot. A second, curated flavor lives in Settings, and both share a combined ten-tag cap. Tag manager (curated tag library) The tag manager is a curated library of session tags that live in their own database — separate from the free-form per-session tag strings stored on each sessions row. Free-form tags (described in session-tags.md) are short strings you type into the picker on the spot; library tags are entries you create once in Settings, optionally scoped to a subset of projects, then re-apply to any session that fits. A tag can also be ANCHORED, rendering as its own collapsible section in the sidebar. Tags — Browse Sessions by Tag The Tags virtual project is a sidebar entry that lets you browse every Claude Code session by the tag attached to it. It follows the standard three-pane project layout: the leftmost project sidebar (where you click Tags to enter), a tag-list sidebar in the middle column at the same width as a normal project's session sidebar, and the main panel on the right showing every session carrying the tag you select, grouped by project. Tree diagrams in chat Agents draw hierarchies with characters, and the diagram only survives while the box on screen is wider than its widest line, so on a phone it reads as noise. A tree block in a chat message is read back into an actual hierarchy and drawn as a real tree, with labels that wrap inside their own node, folding branches, and depth shown by indentation rather than character padding. Turn duration (how long Claude took to respond, shown on every turn bubble) Every Claude turn in an Omniscio chat carries a small duration label: a live ticker while the turn runs, a frozen took badge once it settles, and a rolled-up duration on the activity bubble beside each compact divider. Covers where each one appears, why the inter-compaction figure excludes your idle gaps, how a paused turn inflates the number, and how to read each label. Tweet cards (any X/Twitter link renders the tweet, everywhere) When an X or Twitter tweet link appears anywhere in Omniscio, the app shows a screenshot card of the tweet itself right below the message, no matter how the link got there. Each tweet is rendered once and cached, bursts are capped, failures quietly fall back to the plain link, and only the tweet's numeric id is ever sent to the rendering service. Use Quick Responses (AI suggestions + snippets) Quick Responses are the two fast-reply surfaces in Omniscio: AI suggestion chips, three generated short replies that appear above your input whenever an agent asks a question, and Snippets, your own saved canned responses with usage tracking and sort order. A Zap button sends a one-click preset. All three work in sessions, SMS and email. Verdict results in the transcript (readable result cards) When an agent asks Omniscio's test service to run its tests, the answer arrives in that agent's transcript as a message written for a machine, where the one fact a person wants has the same visual weight as everything else. These arrivals now render as a result card: the outcome in plain words with a matching colour, the numbers that matter, the failing files, and how long it took. Waiting detector The waiting detector decides whether an agent that just finished a turn really needs you or merely described waiting on something it started. This page covers how it holds a session Running instead of dropping it in your inbox, the tag you see in chat, the settings and check-ins around it, how it behaves on every engine, and what it deliberately never suppresses. Waiting detector (part 2) Part 2 of the Waiting detector page: what happens when a silent wait reaches its cap, how an agent-stated ETA becomes the wait window, the Keep waiting and Check in buttons, the automatic check-ins on both sides of your inbox, the inbox right-click menu, the hidden confirmation probe for ambiguous waits and the Plain Speak Wait card. Waiting detector (part 3) Part 3 of the Waiting detector page: everything the detector deliberately never does, the optional AI second opinion that double-checks the pattern match, and the first half of the technical reference for anyone working on the code — the single chokepoint every needs-you flip passes through and the pattern library behind it. Waiting detector (part 4) Part 4 of the Waiting detector page: the ScheduleWakeup check-in mode an agent can opt into, then the rest of the technical reference — the three rejection guards, the live-corpus validation, every setting and kill switch, the subagent defense-in-depth gate and the idempotency and race protections. Which hub an alert-started session uses Every other way of starting a session uses the hub you are already in. The **Start session** button on an alert is the one exception: it resolves a hub of its own — the repo the alert names first, then your *Start alert sessions in* setting, then an automatic Omniscio checkout, and the built-in Claude hub last. Woken No-Op Sentinel When a background event re-wakes a session that had already finished, Omniscio teaches every engine to answer with one canonical sentence, detects it as a whole-message match, hard-hides the throwaway reply and does not re-ping you, so a pointless background wake becomes a true no-op. Work through interrupted sessions — auto-advance When you act on a session in the sidebar's Interrupted section, Omniscio opens the next one for you, so you can clear the pile without hunting for the next row — and it never crosses into another section. Working through paused sessions (auto-advance removed 2026-06-19) The per-status "work through the paused pile" auto-advance was removed on 2026-06-19. Acting on a paused session (archive, snooze or reply) now advances exactly like every other status, through one shared resolver with no per-status special-casing. Your own scrolling always wins (the reader owns the view) Once you move a session's transcript yourself — with a finger, the mouse wheel, the keyboard, the scrollbar or any scroll control — nothing automatic moves it again while you stay on that session, except keeping you at the very bottom while you sit there. Opening the session fresh, sending a message, or tapping jump-to-latest hands the view back. Phone and desktop alike. Agents & Crews Agent Crews (mission teams of agents that report to a lead) Agent Crews let any agent session join a mission team as its lead or a member. The app names it, gives the lead a heartbeat, sends every member turn to the lead instead of your inbox, shows you the conversation with the lead — which asks you things through inbox cards — runs helpers on a lean prompt and archives them when done, gives the crew one private board, keeps its deliverables, tasks and statuses in the app, and shows every crew on a live dashboard in the Overseers hub. Agent Firewall (screen tool output for prompt injection) A screening layer that inspects what a tool hands back to an AI agent — a web page, a file, a command's output — and flags hidden instruction text, then warns the agent, raises an inbox alert, or blocks the content depending on the enforcement level. Agent Messages The Agent Messages panel puts every message your agents sent each other into one list, and gives the ones that never arrived a tab of their own. It is read-only, raises no alerts of its own, and ships switched off until you turn it on in Settings. Agent Permission Level (choose how much agents can do before Omniscio asks you) One setting that decides how much freedom your agents get before Omniscio stops and asks you: Read-only, Guarded, Autonomous, or Full trust. You choose it during onboarding or any time in Settings, and it applies to your Claude Code sessions immediately, to your external coding agents (Kimi Code, Hermes, DSH), and to Codex sessions you start afterwards. Block agent messages How to stop other agents messaging one session permanently, so the traffic no longer wakes it or costs tokens. Covers how Block differs from Hide, everything that still gets through, what happens to a message that is refused, who is allowed to turn it on, and the command-line route. Boards — one read-only window onto every conversation your agents have Boards is a read-only panel showing the shared agent board: every channel in use and the messages posted to it, grouped by channel. This page covers where the row lives in the sidebar, the two-column layout of channel list and post pane, its empty, loading and error states, why it is desktop-only, the per-repository board every project gets, who is told when something is posted and how the three read levels work, and the files behind it. Cross-Session Messaging — Agent-to-Agent HTTP Cross-session messaging lets one Claude session in Omniscio send a message to another session in the same app — waking a stopped one, holding a message for a busy one, or telling every agent working the same repository something at once. Cross-Session Messaging — Agent-to-Agent HTTP (part 2) Part 2 of the cross-session messaging page: what happens to a message that cannot be delivered, how an arrival is drawn for the receiver and for the sender, who the overseer is right now, and the controls that quiet or unfold an exchange. Cross-Session Messaging — Agent-to-Agent HTTP (part 3) Part 3 of the cross-session messaging page: the delivery limits and retry safety, the plain-language boundaries of the feature, and the full request and response reference with the code that implements it. Custom model instructions (write your own instructions for a specific engine or model) Free-text instructions you write once and Omniscio adds automatically to the prompts of a chosen AI engine or a chosen model — so Codex can be steered one way and Claude another without retyping the guidance into every session. Covers where the panel lives, how rules and their targets combine, and which channel carries the text to each engine. Default Agent Instructions The curated instruction presets Omniscio injects into every session it spawns — what each preset does, where the settings live, how the text reaches every engine through the system-prompt bundle, and why the attribution preset writes to the local settings file. Hooks The panel for managing the Claude Code hooks Omniscio composes into every session it spawns — both the built-in ones it ships and your own — kept in Omniscio's own database rather than written into your global Claude config. Interrupting or waking another session (the two confirmed escalations) Two opt-in escalations when one agent messages another: kill a busy target's turn instead of waiting for it to finish, and wake a paused, archived, ended or errored target instead of being refused. Both are off by default because both have real consequences, and both must be asked for explicitly. Also the per-target circuit breaker, which stops a sender looping against one target after three terminal refusals and tells it to re-route instead. Replying to a message from another agent When another agent messages you, a normal chat reply never reaches them — your answer lands in your own transcript and the sender simply waits. This page is how to send the reply back over the local command line instead, covering the retry id, what to do when the target is busy, and how to spot a message that was copied to several sessions at once. Swarms — give a goal to a lead agent that runs a team of workers A swarm is a lead AI session plus a pool of worker sessions, all driving at one goal you write in plain English. You give it a goal, a worker cap and a dollar budget; it works out what needs doing and puts agents on it. Unlike everything else in Omniscio that runs several agents, you do not author the steps: Swarms are in development and off by default. Turn them on at Settings → Lab (swarmEnabled). Missions & Overseers Find Ways to Integrate AI Into Your Work and Life A guided mission that starts a real AI-workflow interview session: the agent asks lightweight questions about your work, finds specific AI opportunities, then produces practical implementation ideas and example prompts. Launched from the mission picker or the real-agent onboarding tour. Missions hub (pick a guided job and an AI does it for you) A permanent, chat-first home for Omniscio's guided do-it-for-me missions: pick one and an AI does the whole task in a normal session, driving the app's own tools. A sidebar entry nested under Prompt Tools, reachable from its row or the Missions toolbar button, and separate from the onboarding surfaces that were retired. Overseer worker titles (how a spawned worker is named) When an Overseer spawns a worker session, that session is titled in one fixed shape — a category, the role, a never-repeating number and the task — so you can tell at a glance which Overseer started it and what it is for. Covers what the parts mean and why the number never restarts. Overseers (the hub that runs your agent fleets) An Overseer is an always-on coordinator that watches sessions, hands out work, and tells you when something needs you — and its swarms are the groups of workers it runs. This is the overview: what one is, what the hub looks like on desktop and on a phone, and how you create one. Overseers — running one (part 2) Part 2 of the Overseers page: keeping one working — how it wakes up, stays alive across restarts, and watches for stalls; how to copy one or run several; how to point one at specific sessions and where the answers it collects end up; and how to ask it something or tell it it got something wrong. Overseers — the work, the settings and the internals (part 3) Part 3 of the Overseers page: the kinds of work an Overseer can hand out and the jobs it may start on its own, what it notices without being told, how agents hand notes to each other, the spend circuit breaker and the settings that govern it all — plus the internals for anyone working on it. Project Kickstart (second guided mission) Project Kickstart is Omniscio's second coached mission: it turns an idea into a real starting point on disk — a new folder holding a plain-language plan and one starter file. Covers where it is launched from, the six decision-card beats, what each type produces, the undo guarantee, and v1 limits. Set Up My Daily Briefing (guided digest-enable mission) A short guided mission that turns on Omniscio's built-in Daily Digest for a non-programmer and lets them choose the morning hour it arrives, so a once-a-day briefing of their Omniscio, email, calendar, and chat activity lands in the Briefings tab. Teach Omniscio About Me (guided personalization mission) A short coached mission that interviews the user for a few facts about themselves — what to call them, their world, how they like answers, and what they're working on — then saves a tidy "About me" block into Omniscio's global Custom Agent Instructions (defaultAgentInstructionsCustom). Because that field is injected into every Claude session via .claude/amc-instructions.md, every future session starts out already knowing the user. Windows-only in v1. Fully reversible. Team Chat reply watch — an agent hears back from a person When an agent sends a Team Chat message on your behalf, it now hears the answer. A later message from a real person in that conversation or thread is delivered straight into the agent session that asked — framed as that person's words, naming who wrote it and how to answer — once per message, never twice, and never into a session you have closed. You can see the active watches and stop any of them. Inbox & Notifications 59 Everything waiting on you, and how Omniscio tells you about it.
Inbox Approvals Hub The Approvals hub is a sidebar hub that lists the pending CLI approvals — changes an agent or external tool asked Omniscio to make through the local control server — giving them a home on desktop and mobile instead of living only in the Inbox. Connection Request Inbox Items A Team Chat connection request arrives as a row in the unified inbox, even for someone who has never switched Team Chat on, so a request is never missed. Covers where the row and its detail pane sit, what Accept, Decline and Dismiss each do, and why dismissing does not answer the request. Empty-inbox discovery tip The one-time card Omniscio shows in the Inbox the first time it has genuinely piled up, explaining that archiving is how you clear it — with a button that archives the card itself so you see the gesture happen. Inbox agents An agent that owns your Needs-You inbox. It checks everything waiting every five minutes, settles or routes what it can so the agents blocked on a reply keep working while you are away, and reports back in a short digest on a cadence you choose. It learns your sorting rules as you correct it and keeps them where you can read and change them. Off until you start one. Inbox Attention Analytics — how your inbox triage behaves A built-in, entirely local report on how you work your Omniscio inbox — how long items wait before you clear them, how deep the backlog gets, how often you open an item and move on, and how long you look at each one. It is the inbox counterpart to UI Usage tracking, and nothing it measures ever leaves your machine. Inbox backlog nudge (a reminder to archive when sessions pile up) One gentle inbox card a day, raised when a real backlog of sessions needing you has been waiting for days while you are otherwise working, suggesting you archive what you are done with. It uses no AI, costs nothing, fires on age as well as count, and can be turned off from the card itself or in Settings. Inbox overseer gate (hold new items for one Overseer to screen) The inbox overseer gate lets one Overseer screen your inbox: a new thing that needs you is held briefly and shown to that Overseer first, which either lets it through, puts it away with a reason, or lets the clock run out so it arrives anyway. This page covers the switch, how long a hold may last, what is never held, where you see and release what is being held, and what the Overseer can do itself. Inbox overview The Inbox is Omniscio's cross-project triage view: one place listing everything waiting on you, whatever channel it arrived by — sessions, messages, approvals. It is not a separate mailbox but a live aggregation, so triaging from here is the same as triaging each item where it lives. Inbox Pilot (an AI that triages your inbox for you) An AI that reads the sessions waiting for you in your inbox and handles the routine ones — replying, archiving, escalating — so you only see the items that genuinely need a person. It can be run in observe-only mode first, is bounded by cost and pause controls, and is per-session opt-in. Inbox Pilot (an AI that triages your inbox for you) (part 2) Part 2 of the Inbox Pilot page: running it in observe-only mode before trusting it, turning it on for new sessions or in bulk, the pill that shows it took a look, the decisions log, what it does with your data, and the limits worth knowing. Inbox Rules Engine (auto-suppress / archive / snooze matching alerts) Rules that act on inbox items automatically, so noisy or known-benign alerts never pile up unread. A rule pairs a match with one action — suppress silently, archive, or snooze until a time you choose — and the highest-priority match fires once, before the item ever surfaces. Off by default. Instant SMS taps in the inbox (preload) On the phone / web view, tapping an SMS conversation in your inbox used to show grey "loading" placeholders (a shimmer) for about half a second to a second before the messages appeared. That wait is a real network round-trip: your phone reaches Omniscio over the Tailscale tunnel, so every tap fetches the conversation across the network. Merge-tooling approval card The merge-tooling approval card: the one inbox card that asks you to approve a change to the small helper programs the app runs while it merges your branches. Explains why the app only runs a version a person approved, what the card shows, and what Approve and Decline each do. Settings patch approval (review CLI / AI proposed settings changes before they apply) When something outside the main UI — usually an AI agent — tries to change one of your Omniscio settings, the change does not apply immediately. It waits as a pending approval in your inbox, shown as a plain-language diff of exactly what the patch would do, with the setting's own description and its before and after. You approve to apply it or reject to discard it. Snooze any inbox item Every row in the Omniscio Inbox is snoozable — sessions, SMS conversations, Telegram chats, fired alarms, cron approvals, automation approvals, recipe approvals, recipe authoring approvals, CLI Pending Actions, RSS articles, RSS digests, drip campaigns, the Briefings digest, and weekly summaries. Snoozing temporarily hides the row from the Inbox and from every other "needs you" / sidebar / list filter until a time you pick, then the row reappears automatically. Snooze everything else (keep a few things, hide the rest for a while) Snooze everything else is snooze turned around. A normal snooze asks you to pick what to HIDE. This asks you to pick what to KEEP — then it hides everything else until a time you choose, and brings it all back on its own. It exists because the normal way round gets harder exactly when you need it most. If your inbox has 23 things in it and you want to concentrate on 3, the ordinary bulk snooze makes you select the other 20. The noisier the inbox, the more work it is to quiet it. Spam filter (texts, agent email, Help Desk + calendar) The spam filter keeps junk from strangers out of your texts, your agent's email, the Help Desk support address and your calendar invites. An AI check judges each stranger's message, caught messages wait in one Spam list in the Inbox, nothing is deleted on a guess, and your "Not spam" and "Mark as spam" clicks teach it. It is in development and off until you turn it on. Stuck-task helper (get unstuck when you keep snoozing the same thing) The Stuck-task helper notices when you keep putting off the same thing and offers a hand. When you snooze the same session, email, pull request, or message three times without ever working on it in between, Omniscio quietly drops one dismissible card in your Inbox: "Want a hand getting unstuck?" Accept it and Omniscio starts a short AI chat that's already been told what you've been avoiding — it interviews you to figure out what's actually in the way (you're not sure where to start, it feels... The inbox review clock Your inbox review no longer depends on an agent remembering to run it. Omniscio wakes your inbox overseer on schedule, hands it the whole picture in one call, then watches what becomes of the review: an agent that finishes it keeps the round, Omniscio runs the review itself on a low-cost model when the agent is silent, and a round nobody could run is closed as interrupted. Every item seen gets a recorded outcome and a reason, and standing in can only ever leave an item for you. Trickle back (staggered snooze — bring a pile back a few at a time) Trickle back takes a pile of selected inbox items or email threads and brings them back a few at a time on a cadence instead of all at once. You choose how many return at a time and the pace, either a simple interval or a clock schedule. It is a variant of snooze: a trickle returns its items in waves so you are never buried again. Trim unused snooze times (personalized snooze menu) The snooze menu offers a standard list of durations (an hour, this evening, tomorrow, next week, and so on). Trim unused snooze times quietly removes the durations you never pick, so the menu shows only the choices you actually use — the rest stay in their normal order, just shorter. It is a small decluttering convenience: it does not change what snooze does, only which options the menu presents to you. When a text leaves your inbox An incoming text puts its conversation in your Inbox and marks it unread. Opening that conversation now clears it from the Inbox at the same moment — reading a text is the same act as dealing with it. The app only ever removes a thread you actually opened: the automatic selections it makes for you leave the Inbox untouched. Alerts & Notifications Briefings (AI briefing of what happened) Briefings — the sidebar's Daily Digest — is the once-a-day AI-written recap of everything that happened across your sessions, mail, calendar, SMS, Slack, Telegram, RSS and GitHub. Covers what it includes, how to tune its hour and sources, follow-up questions, alert triage, hidden sections, the per-day usage ledger behind its token and cost figures, and excluded projects. Coaching Tips (contextual hints and shortcut nudges) Context-aware nudges that Omniscio raises as dismissible inbox cards when you are doing something the hard way — the five opt-in categories, the frequency caps, how to mark a tip learned, snooze, or dismiss it, and why the whole feature ships off by default. Cron failure alerts What happens when a scheduled job fails for good: the Windows toast, the persistent inbox card that stays until you acknowledge, snooze, or the job recovers, the detail pane's Fix with AI / Pause / View actions, and the settings and silence rules that govern them. Daily Brief (what shipped, what is blocked) One calm inbox card a day telling you what the agents shipped, what could not merge, what is held waiting for you, and how many dev-pipeline approvals Omniscio cleared on your behalf. This page covers what the Daily Brief is, how it differs from the AI-written Daily Digest, how to turn it on, and what each line of the card actually means. Declutter unused quick replies (remove one you've stopped using) How Omniscio notices a saved quick reply you have stopped using and offers to remove it in one tap — where the card appears, what counts as unused, what it never flags, and the setting that turns it off. Developer Broadcasts A way for Omniscio's developers and admins to push an inbox card to users' apps — everyone, or a group — without shipping a new build. What you can put in one, how delivery and read receipts work, and how to turn it off. Doc Token Alerts (heads-up when agent-instructions grow large) The amber inbox card Omniscio raises when a project's always-loaded agent documents grow past a token threshold — what it counts, the three ways to act on it, and how to change or silence the threshold. Feature Discovery Nudges The gentle, roughly-daily inbox card that pitches one feature you have never used — how a nudge opens in the main panel, how the "Start using it" button works, how nudges rotate and auto-stop, and how to see what they have achieved. Feedback channel opt-out (silence outbound email per install) Three switches that silence the email leg of Omniscio's outbound reports for your install alone — what each one governs, what it deliberately does not change, and why the local on-disk copy of a report always survives. Inbox alerts (how an agent gets your attention) The cards an agent, a cron job or the app itself raises into your inbox to ask for something: what an alert looks like, which types exist, how you answer one, and how duplicates and repeat problems are kept from flooding you. How they are authored and what happens after you act continue on the parts. Inbox alerts (how an agent gets your attention) (part 2) Part 2 of the Inbox alerts page: how an alert is actually authored — adding a new alert type, the route an agent posts one through, the text handling it has to survive, and the app-generated cards that arrive with their own custom actions attached. Inbox alerts (how an agent gets your attention) (part 3) Part 3 of the Inbox alerts page: the Alerts screen and an alert's life after it arrives — the cooldown that keeps a dismissed problem quiet, the recently-dismissed view, starting a session from an alert, automatic remediation, provenance, and the once-a-day digest. Lingering emergency settings-revert alerts When an AI agent flips one of your settings off as an emergency workaround, nothing used to switch it back on once the bug was fixed. Omniscio now audits those approved agent changes in the background, and when it finds one that is three or more days old and still overriding the app default, it raises a single inbox card. The card quotes the agent's original reason, and its button jumps straight to that setting in Settings. MCP Overhead alert (when your MCP servers cost too much) An amber inbox card that fires when the MCP servers configured for your sessions add up to too much per-request overhead — too many tool definitions in context, too much memory while sessions run, or servers you never actually use. Covers the three independent trip conditions and the one global rule behind them. Notification History (past toasts and dismissed alerts) The one place to look back at notifications you may have missed, with two tabs — toasts that already came and went, and the inbox alerts you dismissed. Opens from a single row in the sidebar's System group, works the same on desktop and mobile, and every row expands in place rather than opening a screen. Notifications & Silence (when Omniscio is allowed to interrupt you) Everything that decides when Omniscio is allowed to pull you back in — OS toasts and their three-way behavior, the chimes each alert type plays, mobile push — and every rule that quiets them again: per-session dedup, a global cap on how often any sound plays, and the moments the app knows you are already looking. Repeated-phrase quick-reply nudge (save a phrase you keep typing) How Omniscio spots a short phrase you keep typing and offers, once, to save it as a quick reply — the repeat threshold, what it counts and deliberately skips, and the settings that control it. Skill Budget Alert (per-engine skill overload) When Omniscio syncs your skills into other AI CLIs (Codex, Gemini, Cursor), a strict-loader engine like Codex can receive more skills than it can handle and warns "skills over length". This feature notices that and lets you decide what to hold back — it never cuts skills on its own. It is the fan-out sibling of the Skill Overhead alert: that one watches the skills loaded into your Claude context; this one watches how many skills get copied to each other engine. Skill Overhead Alert (Skills Bloat checker) Amber inbox card that fires when the Claude Code skills you have installed add up to too much always-on overhead. It is the skills analogue of the MCP Overhead Alert: same inbox-card + snooze + Settings-limit shape, but global (skills config is not per-project in Omniscio — every installed skill loads into every conversation, so there is exactly one rule) and with two independent trip legs instead of MCP's three (skills hold no RAM while idle, so the RAM leg is dropped). Weekday Reflection Nudge (a timely once-a-weekday "could an agent do this?" inbox card) A gentle once-per-weekday inbox card that nudges you to step back and ask whether an Omniscio agent or automation could take some of this off your plate. It never lands while you are away, and the wording rotates daily without ever sounding accusatory. Showing it costs nothing. It is still in development and off by default. Weekly Summary (Monday-morning recap, folded into Daily Digest) Weekly Summary is the Monday-morning AI recap of how you used Omniscio over the just-completed week, plus personalized suggestions. It ships as a briefing inside Daily Digest, is on by default, and is read from the Briefings sidebar entry or the two inbox cards. Weekly Summary (Monday-morning recap, folded into Daily Digest) (part 2) Part 2 of the Weekly Summary page: the feature recommendation rules, the Build it now buttons a suggestion can carry, the first-launch backfill and its past-due gate, the nominal AI usage figure, the phrase exclusion, the commit tally, the cost cap, the Workflow Coach retirement and the stored tables and settings keys. Workflow Coach (RETIRED 2026-05-14 — subsumed by Weekly Summary) The Workflow Coach feature was retired on 2026-05-14 and subsumed by the Weekly Summary, because the two overlapped in purpose: a weekly, privacy-preserving look at how you use Omniscio with AI-generated suggestions. The sidebar panel and weekly inbox alert are gone, the settings moved, and the six detectors now feed the Weekly Summary instead. You don't need to watch a running agent (the in-session hint + weekly leverage card) Two gentle on-by-device nudges that help you stop babysitting a running agent, because it comes to your inbox the moment it is done. One is a small reminder that appears inside a session when you have been watching for a while; the other is an occasional inbox card tallying the time you spent watching and suggesting you run more agents in parallel. Zap coach (a nudge when you type something you could one-tap) A local, no-AI background check that notices when you hand-type a phrase you could have sent in one tap through your Zap preset, a saved quick reply, or the Please continue button, and drops a single gentle inbox card naming the shortcut. It is off by default, costs nothing, runs entirely on your machine, and never tips you about the same phrase twice. Messaging Channels Approved senders for message channels (who can auto-start a session) A per-channel allowlist of who is allowed to auto-start paid work — a session, a reply to the sender, or a workflow — from an inbound text, Slack message, or Telegram message. An unapproved sender still reaches your inbox and still runs your free rules; only the paid parts are gated, and an empty list means nobody. Copy a verification code from a text Omniscio reads verification codes out of your texts so you never have to retype one. A text that arrives carrying a code raises a toast naming the code with a one-click Copy, and a conversation that is nothing but codes (a bank, a delivery service, an app that texts you a login code) gets a large "Most recent code" panel pinned at the top of the thread. Every code it offers is one you RECEIVED — a code you typed yourself is never offered. Emoji and GIFs in SMS The SMS composer has two extra buttons next to the picture button — an emoji picker and a GIF picker — in both the New SMS box and the reply box inside a conversation. Emoji — opens a searchable panel (with a "Recent" row of your go-to emoji); tap one and it drops into your message right where the cursor is. GIF — opens GIF search powered by Klipy (a free GIF service used by apps like WhatsApp and Discord); trending GIFs show when you open it, type to search, tap one to attach it. Quick SMS (Control Space → SMS tab) How to fire off a text without leaving the keyboard, from the SMS tab of the Quick Launch composer — picking a recipient, the short undoable delay before it sends, and the SMS Favorites list that sits behind it. RSS feed integration (read RSS/Atom feeds in your inbox) Omniscio's RSS integration polls RSS/Atom feed URLs you configure and surfaces each new entry as its own row in the unified inbox, alongside Gmail, SMS, Slack, Telegram and webhooks. It is read-only, and it can also produce an AI-curated daily digest of your feeds. Schedule a text (SMS Send Later) SMS Send Later lets you queue an outbound text and have Omniscio send it automatically at a time you choose — once, or on a recurring schedule (daily, weekdays, weekends, or specific days of the week). It is the texting counterpart of the agent Send Later feature, but instead of deferring a reply into a Claude session it sends a real SMS through your selected SMS provider (Pushbullet or native), exactly as if you had typed it and pressed Send yourself. Send and receive pictures in SMS (MMS) Omniscio can send and receive pictures in a text over the Pushbullet bridge (Omniscio mirrors your Android phone's texts through Pushbullet). Pictures ride on your real phone number — this is not a separate texting service like Twilio, so to the other person a picture looks exactly like one you sent or received yourself. Set up SMS integration Omniscio can send and receive text messages using your own Android phone — no Twilio and no separate SMS number. Once connected, Omniscio notifies you by SMS when a session needs your attention, and anything you reply from your phone appears inside the app. Two ways to connect: Pushbullet, which mirrors texts through its cloud and handles pictures; or Native, which goes straight over your own carrier. SMS short-code name inference (auto-fill descriptive contact names) When an SMS conversation comes from a short code (a 3–6 digit sender like 62438 or 22395, no + prefix), Omniscio auto-fills a descriptive contact name for it — 62438 becomes "eBay security code", 22395 (which mixes Amazon, eBay, and Walmart codes) becomes "Verification codes (multi-brand)". Telegram Bot channel (talk to your agents from Telegram) The Telegram Bot channel lets you chat with your Omniscio agents from Telegram. You create a bot with BotFather, paste its token into Omniscio, and then every message you send that bot goes to an agent session, with the agent's final answer posted back in the same Telegram chat. It is separate from the Telegram integration and never touches your Telegram account. Webhooks (receive HTTP POSTs into your inbox) Webhooks run a local HTTP server that accepts incoming POST requests and turns each one into a row in the unified inbox, alongside Gmail, SMS, Slack, Telegram and RSS. Every source has its own URL path and bearer token, the channel is receive-only, and the page covers enabling it, adding and testing a source, the full set of response codes, and duplicate suppression. Where the unread SMS texts start (the "N new messages" divider) When you open an SMS conversation that has texts you have not read, Omniscio draws a small accent marker on an accent line — "3 new messages" — right above the first text you have not read yet, so a long thread tells you at a glance where you left off. It is the same marker Team Chat uses. Projects & Tasks 95 The work you keep in Omniscio — folders, to-dos, boards, notes and your own writing.
Projects Add a project Adding a folder as a project in Omniscio: the Add Project dialog's Quick Create, Browse Existing and GitHub tabs, picking a color and a sidebar group, the background icon detection, the name rules, and the IPC and GitHub plumbing that sits behind the dialog. Agent Instructions Sync A per-project background job that keeps your canonical CLAUDE.md (or AGENTS.md) mirrored to the other filenames agents look for, with drift detection, a conflict resolver, an inbox alert, and two ways of handling the copies in git. Auto Context — auto-inject files like Claude.ai Projects Auto Context is a per-project knowledge folder: drop reference files into a project's .claude/docs/ and Omniscio folds them into every new session's first message. Covers the always and rag buckets, the sidebar group, the doc ordering, the first-message chip, and every way to add files — drag-and-drop, the modal, or an external agent. Auto Context — auto-inject files like Claude.ai Projects (part 2) The second half of the Auto Context page: flagging a doc as an authoritative project instruction, letting an external agent add or remove files through CLI Control with an inbox approval, and the machinery behind the injection — the builders, the per-engine first-message paths, the budget and the kill switch. Auto Context — selection mode + bulk actions A multi-select mode on the Auto Context sidebar panel: pick several always-inject or on-demand docs at once, then copy their contents to the clipboard or delete them in one confirmation. Covers the Select toggle, the keyboard shortcuts, which docs participate, and the bulk-copy cap. Clean Room (spawn a vanilla AI with none of your customizations — a control group) A built-in project whose sessions spawn a deliberately vanilla AI — no global rules, memory, skills, project instructions, connected tools, or Omniscio notes — so you can see what the engine does from a blank slate, and how each session is fenced to its own private room. Default Claude Project (chat without software context) The pre-created Claude project that lives at ~/Claude — Omniscio's general-purpose workspace for anything not tied to a codebase: how to find it, how to start a chat in it, and what unpinning, deleting or moving it does. Delete a project How to remove a project from the Omniscio sidebar: the ordinary removal that leaves your folder untouched on disk, and the separate permanent erase that deletes the folder too — plainly irreversible, with no undo. Deploy Profiles Per-project deploy instructions that Omniscio hands the agent, so it knows how this particular project ships instead of guessing — what a profile holds, how to write one in the project settings, and how it is read back. Edit a project Everything about a project you can change after it exists — its label, colour, emoji or image, pin state, sidebar group, icon suggestions and session isolation — plus the one thing you cannot edit in place, its folder path. File Explorer (browse a project's files) The collapsible sidebar pane that shows a project's folder on disk without leaving Omniscio — expand and navigate the tree, open and preview files, and switch to the Changes tab for the git diff review surface. Project Notes (markdown scratchpad per project) Project Notes is a private markdown scratchpad attached to each project in the sidebar — deploy steps, credential reminders, design sketches — stored as a local markdown file with manual save, a 500,000-character cap and no sync, and deliberately never shown to Claude during a session. Reorder projects You control the order of the projects sidebar yourself: drag rows up and down, group them under collapsible dividers, pin the most important ones to a fixed top section, and collapse groups you do not want to see right now. The order is remembered per account and survives restarts, and ungrouped projects and dividers share one top-level sequence. System Instructions (Auto Context sub-group) One-click access to the four AI-context files that shape a project's sessions — the project's own CLAUDE.md and MEMORY.md and their global counterparts — as the System Instructions sub-group of the Auto Context panel. Covers where the group sits, what each row shows, what clicking does, and why these rows are never bulk-selectable. Tasks & Boards Arij (issue tracker) Omniscio's own Jira-style issue tracker — boards, backlog, sprints, issues, docs and automation — where an issue can launch and track a real agent session. Covers opening it from the sidebar, orgs, projects and members, the Omniscio account it signs in with, and the notification inbox. MC Chat Bot (Mission Control alerts posted into Team Chat) A bot that posts Mission Control project-health alerts — at-risk items, cascading delays, pattern detections, health drops and periodic digests — into a Team Chat channel, so your team sees project risk without opening Mission Control. Off by default; covers the five alert categories and the settings that configure it. MC Dashboards (Mission Control cross-board dashboards) A workspace-scoped dashboard that pulls data from several Mission Control boards into one view: a drag-and-resize grid of widgets, a filter toolbar that scopes every widget at once, and export. Covers creating one, the widget types, and how board access is enforced. MC Templates (save a board as a reusable template) Save a Mission Control board's full structure — its groups, columns, views and automations — as a reusable template, then deploy it into a new board. Covers saving one, the two-level template picker, the deploy preview that lets you switch off cost-bearing automations first, and editing a saved template. Mission Control (the project management system built into the app) Omniscio's own project management system, opening as a sidebar project like any other: boards of typed columns, a set of alternate views over the same data, and the cross-board surfaces around them. This is the overview page — the detail sits on the parts below it. Mission Control — automation and the internals (part 5) Part 5 of the Mission Control page, for anyone working on it: the board automation engine and the events it fires on, the bridge that lets the app's own automation react to a board event, the workflow-engine nodes that poll a board, where the local data mirror lives, and the command surface an agent or script drives. Mission Control — Documents, Spreadsheets & Forms An in-development area, on by default, that adds documents, spreadsheets, and forms to Mission Control workspaces, each scoped to one workspace with the same never-really-deleted storage as the rest of the product. One switch reveals all three together — navigation sections, list and detail pages, six routes, and ten ready-made templates. Mission Control — templates, calendars, meetings and notifications (part 4) Part 4 of the Mission Control page: how boards connect to the rest of the app and to the world outside it — board and recipe templates, the calendar overlay, session linking, the meeting bridge, voice actions, search and knowledge, outbound webhooks, and the notification intelligence that decides what is worth telling you. Mission Control — the table, the views and the board (part 2) Part 2 of the Mission Control page: the Main Table and everything you can do to it — filters, grouping, the alternate views over the same items, the Kanban board's card actions and swimlanes, finding a teammate, and the per-board preferences you can set. Mission Control — workspaces, people and publishing (part 3) Part 3 of the Mission Control page: workspace membership and accounts, sharing a board or inviting someone into a workspace, the permissions and trust controls around them, and the strategic layer above a board — goals, portfolios, and publishing work to a public web address. Monday (the earlier name of the built-in board app) The original name of the board app now called Mission Control: a self-hosted kanban and project-management tool with a table view, alternate views over the same items, board favorites and grouped search. Kept for history — the feature was renamed, and the current documentation lives on the Mission Control page. Monday — workspaces, sharing and the internals (part 2) Part 2 of the Monday page, kept for history: workspaces, accounts, board sharing and invites, the local data mirror, board templates and the automation engine — the parts carried over unchanged into what is now called Mission Control. My Work A cross-board personal task view that aggregates every item assigned to you across all Mission Control boards, organized into six date buckets that follow your local timezone, with a table view, a calendar view, a search filter and a quick-create modal. Ops Engine (PM Operations Engine) An in-development operational dashboard for the AI Project Manager, built as a virtual project that docks in the main panel and shows workspace health, agent sessions, deadline confidence, risk, velocity, shift handoffs, a decision queue and per-board notification preferences. PM + Team Chat Integration (item mentions, task creation, status updates) An in-development bridge between Mission Control boards and Team Chat: mentioning an item with # in the composer, turning a chat message into a task, announcing item status changes in a channel, and listing an item's linked chat messages. Covers each capability, the gating, and the channels behind it. PM Agent Trust (G13 — agent trust levels and approval gates) Controls what AI agents may do on Mission Control boards. Each board carries one of three trust levels (read-only, status-only, full-access) that caps autonomous action, and every agent action falls into an approval tier that either always needs a human yes, can be toggled, or never needs one. Progressive trust promotes repeat-approved actions, and a three-level undo reverses agent changes. PM Calendar Overlay Shows PM item due dates and sprint timelines on the Google Calendar panel alongside your real calendar events, read from a local mirror rather than fetched from the cloud. Events are color-coded by board, highlighted when overdue, and strictly read-only, with a one-click jump back into Mission Control and two settings that turn the overlay on or narrow it to chosen boards. PM Composability (G3 -- feature interaction validation) An in-development validation engine that governs which of the 65 PM features can be enabled, disabled, and combined. It models features as a directed graph with four edge types, checks interactions at enable time, at runtime, and at disable time, and ships 13 internally consistent presets across lane, engineering, and platform-import categories. PM Cross-Board Intelligence (G2 -- cross-board reasoning infrastructure) Shipped backend infrastructure that aggregates every Mission Control board into materialized snapshots across hot, warm, and cold computation tiers, so higher-level features read pre-computed numbers instead of querying live. It backs template-based natural-language answers, per-board seasonal urgency, lead-time tracking, and three tiers of detected patterns. PM Cross-Entity Links Bidirectional linking between a Mission Control item and any of 14 other kinds of thing in Omniscio — vault notes, mind maps, diagrams, whiteboards, flowcharts, URLs, writer documents, email threads, meetings, and GitHub pull requests, commits, branches and issues. Links are made from either side through a search-and-select picker, shown on the item and on the linked tool, and unlinked rather than destroyed. PM Data Density (earn-the-right-to-speak gatekeeper) An in-development gatekeeper that decides when a Mission Control board has accumulated enough usage data to be spoken to intelligently, tracking six domains separately and holding fresh-start users to higher thresholds than people who imported from another platform. Its behavior-pattern engine watches item history for six recurring habits, algorithmically and at no model cost. PM Discovery Callouts (proactive inbox suggestions) An in-development feature that puts proactive suggestions into your inbox: nudges to connect Mission Control to other Omniscio tools, offers to automate a habit it noticed you repeating, suggestions to turn on sprints or composability, and setup tips for newer users. A background scanner runs every six hours and will not repeat the same suggestion for a week. PM Email Integration (link emails to PM items) An in-development bridge between Mission Control items and all three of Omniscio's email surfaces — the Gmail viewer, Agent Email, and Supermail — letting you link an email thread to an item, turn an email into a new item, and fire workflow automations when a linked item's status changes. Desktop only, behind a double feature gate. PM Feature Profiles (G12 -- per-board progressive disclosure) An in-development progressive disclosure engine that decides which Mission Control features are visible on each board. Every board carries a profile built from one of three presets — Tasks, Projects or Operations — which users can extend per board, and a nudge engine suggests features contextually from board activity, under rate limits and a post-enable cooldown. PM Goals Hierarchical goal tracking for Mission Control workspaces: workspace-scoped objectives that sit above individual boards, each with an owner, a target date and a health status, nestable under a parent goal, with progress derived from linked boards or set by hand, plus comments and a field-level activity feed. PM Import Framework (import data into Mission Control boards) Omniscio's universal import framework converts data from CSV, Excel, Google Sheets, Trello, Jira, Asana, Monday, Linear, Notion, ClickUp and Airtable into local PM boards through one seven-stage pipeline, with conflict strategies, error recovery, job tracking, connected-source refresh and third-party PII consent. PM Intelligence Advisor An advisory intelligence layer for the AI Project Manager: it reasons across all your project management boards and surfaces strategic insights, completion predictions, policy recommendations, and answers to plain-English questions about your projects. PM Notification Preferences Per-board notification preferences for Mission Control: each board can be muted, promoted, configured for email delivery and given a digest frequency, with session-scoped batching, a never-notify-actor filter, and preferences stored locally and synced over the PM push channel. PM Onboarding and Help (preset-scoped checklist + contextual tips) In-development onboarding checklist and contextual help system for PM users: a preset-scoped getting-started checklist, an auto-created Setup Checklist board whose steps spawn an AI PM session, per-feature help popovers, and a unified help panel with escape-to-AI. PM Onboarding Features (import + post-import flows) Three in-development PM onboarding features for users migrating from other platforms: a persona-C import entry screen with multi-select platform cards and auto-applied presets, weekly discovery callouts, and a personalized post-import Home. PM Portfolios Cross-board aggregation for Mission Control: group multiple boards into a named portfolio and see a unified status summary of total items, completed items and a status-bucket breakdown across every linked board. PM Predictive Analytics (G9 -- velocity, forecasts, risk) An in-development predictive analytics layer for PM boards: velocity tracking over a configurable window, Monte Carlo completion forecasts, what-if scenarios with resource, priority and deadline levers, and risk heatmaps with cascading cross-board impact. PM Search Integration PM items are discoverable through the global search alongside sessions, SMS, Slack, Telegram, RSS and webhook results, by both text search and semantic embedding similarity, with a "Use as context" action and session-linked deduplication. PM Session Link Bidirectional linking between a working session and a Mission Control board item: the session header names the item it targets, and the board shows that session's live agent state. Covers where the link appears, how linking and unlinking behave, the IPC channels behind it, and what is not built yet. PM Sprints Sprint management for Mission Control boards: time-boxed iterations with a planned, active and completed lifecycle, item assignment from the board toolbar, velocity stats recorded on completion, and retrospectives generated either from a template or by an AI model. Covers the lifecycle, velocity tracking, and where each part lives. PM Stickers Decorative emoji-style stickers that users place on Mission Control board item cards, with free positioning, rotation and z-ordering on the card surface. Covers the eight built-in images, how adding and removing behave, where the picker appears, and the channels behind it. PM Time Tracker Bridge Connects Mission Control items to the Time Tracker: a Time tab in the item card starts and stops timers and shows the time logged against that item, and the same entry appears on the Time Tracker's Today timeline. Covers the invariants, the data model, and the channels behind it. PM Voice Actions Hands-free project management: say what you want done on a board instead of clicking. Covers the five supported actions, which of them need a live connection, how ordinary phrasing is tolerated, and the swappable speech recognition provider behind it. PM Voting Per-item upvoting on Mission Control board items to signal priority: users toggle their vote and counts are read per board. Covers where the vote badge and button appear, how duplicate toggles are prevented, and the channels and feature flag behind it. PM Webhooks Webhooks in both directions for Mission Control boards. Outbound: subscriptions POST matching board events to callback URLs with an HMAC-SHA256 signature, retried, health-tracked, auto-disabled and dead-lettered. Inbound: a public per-board URL receives external POSTs, with an optional HMAC signature to prove the sender, and folds each one into a board event. Tasks (markdown outliner, in development) Tasks is the next-generation parallel-task outliner for Omniscio, surfaced as a virtual project in the sidebar (label "Tasks"). It is a keyboard-driven tree of to-do rows where each row's text is real markdown: in view mode the row renders the markdown (headings, bold, code, lists); Tasks (markdown outliner, in development) (part 10) Local control server (127.0.0.1:19519), in src/main/services/cli/cli-server-tasks-v2-routes.ts. Every handler runs the shared feature gate (tasksReadAuth / tasksMutationGate → isCliFeatureVisible('tasks-v2')) first and returns 403 with disabled: true when the feature is off — deliberately not a 404, so a caller can tell "feature off" from "id not found". On these routes a 404 means only that the id is missing. Send the plain local time. Do not convert to UTC yourself. Tasks (markdown outliner, in development) (part 11) The key-files reference for Tasks: one row per file that carries a piece of the feature — the migration that creates the tasks table, the main-process service, the IPC handlers, the renderer components, the CLI routes, and the writer that mirrors the task tree to disk — so anyone working on the code can go straight to the right file. Tasks (markdown outliner, in development) (part 12) The strategic half of the Tasks page: the task data model used as a platform — AI ordering and quadrant classification with their provenance, the day-log reflection ritual, snooze-as-signal procrastination detection, planning personalisation, caught task proposals, delegation records, estimation accuracy, the waiting-on people graph, meeting prep, calendar fusion, and the key-files table. Tasks (markdown outliner, in development) (part 2) These command keys are rebindable. Open Settings → Keyboard Shortcuts → Tasks (the tab appears once Tasks is enabled) to remap any of the command-mode actions — snooze, deadline, priority, estimate, waiting on (A), finish, context, launch agent, break-down, command line, insert link (Ctrl/Cmd+K), the focus moves, the one-hand move keys (Shift+W/A/S/D), cut/paste (Ctrl/Cmd+X / Ctrl/Cmd+V), move to project (V), fold/unfold (F), and help. Tasks (markdown outliner, in development) (part 3) Snoozing tucks the task into a "Snoozed" section. Once a task is snoozed into the future, it leaves the list and collects under a single collapsible ⏰ Snoozed (N) row pinned at the bottom — so the list stays focused on what's actually actionable, while the count tells you at a glance what's parked. Tasks (markdown outliner, in development) (part 4) Tasks doesn't just show deadlines and snoozes on the row — it tells you when they come up. A quiet background check (running whenever Tasks is on) raises one Inbox alert the moment: a task's deadline arrives (its due date/time passes), or a snoozed task comes back (its snooze runs out). Tasks (markdown outliner, in development) (part 5) Every task can carry a little more than its text — all optional, all surfaced as small chips on the row only when set (so rows stay clean): An Importance dial — None / Low / Medium / High. Stored as a real column (importance, integer 0–3) so a later "what should I do right now" view can sort on it quickly. (Urgency was removed from the Tasks surface — its urgency column stays but is no longer shown.) A scheduled start date (startat) — "when I plan to start this," separate from the due deadline. Tasks (markdown outliner, in development) (part 6) A dedicated Today surface holds the day's true priorities across every project. It is a non-destructive overlay: a task on Today keeps its home project and listid — Today just flags it (dailyplatedate = the local day). Nothing is moved or copied; at day's end a task simply drops off Today and is still sitting in its project. It behaves like any other project — fully editable. Tasks (markdown outliner, in development) (part 7) Press / (or Ctrl/Cmd+F), or click the header's Find button (🔍), and a slim find bar opens under the header. Type, and the list filters live — in the giant All-Projects view and in a single project alike. Matches (and their parent tasks, so you can see where they live) stay visible; everything else drops away; the matched letters are highlighted in each row. It searches task text and #tags, ignores case and accents ("cafe" finds "Café"), and shows a live "N found" count. Tasks (markdown outliner, in development) (part 8) Any task can host its own Claude agent chat. On a focused outliner row, the "Launch agent" button or the g key runs open-or-launch: If the task already has a chat, it reveals in the Sessions tab and selects it — free, no spawn, no pop-out. If it doesn't, an editable launch-prompt popup opens, showing the visible task message the agent starts from (the task text + its open subtasks) — editable — plus a session-type picker (Custom / Break it down / Work through resistance / Tame the overwhelm /... Tasks (markdown outliner, in development) (part 9) An eighth agent session type, Define & enrich, is the fuller sibling of Clarify & enrich: it not only sharpens each task's wording, it also captures the context behind it. Like Clarify it is project-launch only — it never appears on a single-task launch (the picker filters it via LISTONLYSESSIONTYPES), because its whole job is to walk the list. Launch it from a project's "Launch agent…" menu (the launch popup in list mode) and pick Define & enrich. Tasks (personal outliner — removed) The current keyboard outliner is Tasks — a separate feature with its own table, store, IPC channels, and mirror file. This page is retained only as a redirect, so that existing references to the removed outliner still resolve instead of dangling. What Changed (PM Unified Feed) A cross-board activity timeline showing what changed across all your Mission Control boards, ranked by how much it matters to you, with filters, inline actions and a daily 8 AM digest. Covers how events are scored into tiers, the batch summariser, and the settings behind it. Notes & Writing Decks (AI presentation builder) Decks is the AI presentation builder in Omniscio: describe a deck in plain language, approve the outline it returns, then get editable cards with real layouts, images and speaker notes that you can reorder, present, publish as a web link, or download as PDF, HTML or PowerPoint. Decks (AI presentation builder) (part 2) Part 2 of the Decks page: what each card is told when it is generated, how the topic colour and the 13-layout library work, publishing a deck to a web link, real images and generative art, presenter mode and speaker notes, PDF, HTML and PowerPoint export, and the data model, IPC surface and key files. Export to Word / PDF (Markdown → .docx / .pdf) Turn Markdown anywhere in Omniscio into a Word document or a PDF saved to your computer, from one Export… menu that also holds the existing online-publish actions. Word and PDF are free, local and always available; sharing online and exporting to Google Docs appear only where they already worked. Flowcharts (visual flowchart editor) A standalone visual flowchart editor that opens as a virtual project in the sidebar: draw boxes and connectors on a canvas, organise them into swimlanes, keep a saved library of charts, and describe a chart in words to have an AI build it. Mind Map (build a map with the keyboard or the mouse) A mind map you build inside Omniscio: open it, add nodes with the keyboard or by dragging, edit their text and attach hidden notes, drill into a layer behind a node, and change how the whole map looks. Covers the saved-maps library and where a map goes when you are done with it. Mind Map — notes, layers, AI and the internals (part 2) The second half of the Mind Map page: hidden notes attached to a node, drilling into a layer behind one, themes and layout, importing, exporting and sharing a map, the two AI features (which cost money), how a map behaves under your finger on a phone, and what is still rough. Scratchpads (persistent notes and the quick-capture overlay) Scratchpads is a Chrome-style persistent notes feature inside Omniscio: open the Scratchpads virtual project for a two-pane pad list and editor, or press Ctrl+Shift+S from anywhere for a quick-capture overlay. This page covers what a scratchpad is, the overlay, popping one out into its own window, undo, find and global search. Scratchpads — part 2 (images, formatting, deletion, the CLI and the contracts) The second half of the Scratchpads page: how images, Markdown styling and list hotkeys behave inside a pad, what happens to a pad you delete, backward compatibility with the old plain-text pads, the command-line surface an agent uses, and the architectural contracts anyone touching the code has to respect. Sticky Notes Floating scratch-pad notes that persist across sessions and app restarts. Each note is draggable, resizable, and color-coded. Notes live in a portal overlay above the main UI and are toggled from the toolbar. Writer Studio (standalone writing editor) Writer Studio is Omniscio's standalone writing surface — a full-screen markdown editor with a documents list, writing guidance, a creativity dial and four AI capabilities, plus how the native surface and the AI Writer plugin relate and how the feature is gated on. Writer Studio (standalone writing editor) (part 2) Part 2 of the Writer Studio page: the formatting toolbar, find and replace, the slash menu, document templates, tags, cloning, bulk operations, rename, font and spacing controls, Save as, the editor theme and colour palettes, and the two writer data-model tables. Writer Studio (standalone writing editor) (part 3) Part 3 of the Writer Studio page: the four AI capabilities the editor ships — selection edits and their review flow, options-mode variants, inline ghost-text autocomplete, and the chat assistant with its suggested edits, persisted history and voice context. Writer Studio (standalone writing editor) (part 4) Part 4 of the Writer Studio page: publishing a document to Shares and sending it by email, SMS or Slack, PDF and Word export, Google Docs import and export, local file import, and the Session, Mission Control, KMS vault, Team Chat and PM-tool bridges. Writer Studio (standalone writing editor) (part 5) Part 5 of the Writer Studio page: the complete writer IPC channel reference and the key-files table that maps each part of the feature to the source that implements it. Personal Organisation Bookmarks Bookmarks save URLs, file paths, programs and shell commands, plus sessions, projects and virtual projects, so they are one click away from anywhere in Omniscio. They are managed from the toolbar popover and stored locally only. Bookmarks (part 2) The reference half of the Bookmarks page: the localhost CLI routes with their kill switch, curl examples and response shapes, plus the implementation map of the queries, launcher, favicons, IPC channels and renderer state. Clipboard history A clipboard manager built into the app — a background watcher remembers what you copy on Windows across every program, so an older copy is one shortcut away. Covers where the composer button and the Ctrl+Alt+V pop-up live, how to search, pin, edit and annotate copies, how long they survive a restart, and the privacy safeguards. Coffer (in development) Omniscio's first-party personal-finance tracker: a fully local Mint-style ledger — accounts, transactions, budgets, recurring bills, goals, trends and alerts — with bank sync as an opt-in phase and no cloud backend. Covers what the panel contains today, how to turn it on, and the wiring and remaining phases behind it. Daily Journal Check-In (AI Coaching check-in nudge) An opt-in ritual inside AI Coaching: a gentle inbox nudge inviting you to journal with your coach — a short daily check-in, and a longer weekly review. Covers what the two interviews are, where the nudge time and weekly day are set, how the entries accumulate as versions of one artifact, and how they stay private to coaching. Eligible phrases (browse & save every phrase you repeat) A tab in Quick Replies that lists every phrase you type often, so you can browse them all and save the good ones as quick replies in one click — running entirely on your own machine, with no AI and no cost. Habits A local habit and daily-routine tracker: define habits, log them each day, and watch streaks and completion build across four views. In development and off by default, and entirely local — no cloud, no AI, no cost. Journal (freeform and guided reflection) A personal-reflection space inside Omniscio with two modes: freeform writing in a plain editor with autosave, tags and a reading-time estimate, and a guided Socratic conversation with Claude that follows one of five reflection templates. Entries stay on your machine. Pomodoro The built-in pomodoro timer: focus blocks and breaks, presets, schedules that auto-start at a wall-clock time, and the today/week/by-project stats. Covers where the Pomodoro tab lives, how a run advances and pauses, how it pairs with Focus Mode, and what it deliberately does not do. Quick Replies (manage your snippet library) How to build and manage your library of saved quick replies: the sidebar virtual project that holds quick replies, dividers and folders, the Settings tab that owns the AI-suggestion toggles, what happens when a snippet fires in a session, and the three library-level actions — Export to a JSON file, Share as a public link, and Import — whose merge only ever adds. Records — gamified personal bests (trophy case) Records is a gamified trophy case inside the Stats virtual project. It tracks about 18 personal-best feats — your longest session, priciest day, longest active-day streak, lifetime milestones, and a Peak Parallelism family measuring how many sessions and projects you have run at once. Locked cards show a hint for how to earn them; earned cards show your best, and some records celebrate with confetti. Time Tracker A built-in, off-by-default personal time tracker: press Start to run a live timer and drop a colour-coded block onto the day's timeline, then tag it, attribute it to a tracker project and mark it billable, with reports, CSV export and invoices all derived locally from the time you actually logged. Typing Insights (typing-speed stat) A small passive feature that measures how fast you actually type in the session composer and shows it as a Typing Speed stat in the Statistics view, with your average WPM for the selected range and your peak WPM beneath it. It runs on by default, adds no typing latency, and is fully local with no AI, no internet and no cost. Automation & Scheduling 75 Work Omniscio does on its own — on a schedule, on a trigger, or as a pipeline.
Automations & Recipes Audit Framework (bundled audit skills) The eight bundled skills that run a compaction-resilient codebase audit and turn its findings into commits: what each phase does, how to start one, the system and first-principles overrides, and how a finished report reaches master. Author a recipe by asking AI (CLI authoring) You can author a recipe by describing it to an external AI — Claude.ai, ChatGPT, or any Claude Code session — which drafts, validates and saves it into Omniscio for you. Every recipe saved that way lands in your inbox as pending and refuses to run until you approve it, so a buggy or malicious AI cannot silently execute work on your machine. Auto-replies (auto-respond to a session's own output) Auto-replies — the lightweight half of Omniscio's two auto-response tools: a single-trigger rule that watches a session's finished output and fires exactly one action, the six actions it can take, the loop-safety caps on the billable ones, and how it differs from Automation Rules. Automation Builder The Automation Builder — the in-app assistant that interviews you in plain English and then builds, red-teams, schedules and registers a real automation, the ready-made presets it offers, how it repairs and monitors what it built, and the cost and safety rules around it. Automation Credentials (saved secrets for v2 actions) Automation Credentials — the manager that stores the API tokens, webhook URLs, service-account keys and SMTP passwords that the v2 automation actions resolve at run time, the per-kind secret editors, the high-risk write gate, and the metadata-only CLI routes. Automations & Auto-replies (auto-respond to messages) The two ways Omniscio acts automatically: Auto-replies fire one action when a session's finished output matches a rule, while Automations run a pipeline of up to ten actions over incoming channel messages, with scoping, backtesting and AI reply review. Automations & Auto-replies (auto-respond to messages) (part 2) The second half of the Automations page: the auto-response badge taxonomy that names which automated path produced a message, the activity log's per-step breakdown, and the implementation map of the engines, stores and IPC. Bot Creation Mode Bot Creation Mode turns one sentence about a job into a working bot: a guided interview that asks a couple of questions, shows you the whole bot, and then builds it as a pre-filled project folder with its own sidebar hub, persona, knowledge files and scoped skills. Bug intake safety screen The bug-report safety screen: before Omniscio starts an agent on a bug report sent from inside the app, handed over by a teammate, opened on GitHub or raised by Sentry, a very cheap AI checks everything the agent would read for prompt injection and dangerous requests. What it holds, where held reports wait, and how to start or dismiss one. Bug Report Intake Bug Report Intake is the Omniscio feature that turns incoming bug reports into investigations: testers email a tagged subject, and Omniscio polls Sentry and GitHub for new issues, spawning a Claude session in the matching project. Part 1 covers what the feature is, where its sidebar panel lives, and how email routing and email rules decide where a report goes. Bug Report Intake (part 2) Part 2 of the Bug Report Intake page: the three channels Omniscio watches for you — a connected Gmail inbox, a Sentry organization and GitHub repositories — how to add, test and edit each source, what happens on every poll, and how a source that keeps failing raises a durable alert. Bug Report Intake (part 3) Part 3 of the Bug Report Intake page: what an auto-spawned investigation actually receives — the fenced report, your appended instructions, email attachments and in-app screenshots — how a reply threads back to its session, how repeat reports of one issue collapse into a single investigation, and the safety guarantees behind all of it. Bug Report Intake (part 4) Part 4 of the Bug Report Intake page: the table of decisions you can see when a report does not route, the analytics events intake records, what is deliberately out of scope for the first version, and a map of every file and route behind the feature for readers who have the repository open. MC Automations (the automation engine behind Mission Control boards) The automation layer for Mission Control boards: an event-driven engine that fires action chains when board data changes, a channel bridge that lets a cross-system automation react to a board event, and the workflow-engine nodes that poll a board on a timer. Part of Mission Control Phases 5C and 5H. My Automations (everything you have scheduled, in one list) The one sidebar list of everything you have running on a schedule — the rich automations you built with the Automation Builder and the simple jobs you asked a session to create. What each row tells you (what it does, when it runs, whether it is live) and how this view relates to the raw Cron Jobs screen. Recipe Drafter (describe a workflow, get a validated recipe) The Recipe Drafter turns a natural-language description of an automation into a draft recipe, returned with its reasoning, warnings and any remaining validation errors. The draft is not saved to disk, so the caller decides whether to save it, edit it, or retry. It reads a bundled pattern library, validates the candidate, and can refine its own draft for up to three turns. Sentry Triage Gate A decision step between Omniscio noticing a new Sentry issue and Omniscio spawning an investigation session for it. The triage gate ignores what you muted, consolidates onto an investigation already running, and grades the rest by severity — and by default it does not HOLD: an issue it would otherwise park starts a session instead, so one underlying bug split into many issue IDs still cannot spawn a pile of near-duplicate paid sessions while nothing waits in an inbox for a click. Session Starter Bot Session Starter Bot is a per-channel automation in Team Chat: it watches a channel for messages matching keywords you choose and automatically starts an AI session for each one, in the project you nominate. You configure one bot per channel, and it runs whenever Omniscio is open. It can also post an acknowledgment, or reply back in the channel conversationally as your agent. Silent recipe sessions Some recipes run dozens of times a day in the background — newsletter summarizers, link-archivers, audit cycles — and you don't want to see their sessions in the inbox or sidebar, period. Silent mode hides the orchestrator session from the inbox and the project sidebar for the entire run — including attention statuses like needsyou, error, and stalled — then auto-archives the session inline on clean completion (no "Archive ses…" row in the inbox). Silent really means silent. Trigger Recipes from an Agent (HTTP) The two HTTP endpoints on Omniscio's localhost CLI Control server that let an agent fire a flagged recipe without a human approval click: discovering triggerable recipes, firing one with parameters, the error codes and cooldown rules, and how the per-recipe Agent triggering flag and cost cap work. Use Recipes (multi-step agent workflows) A Recipe is a saved multi-step workflow that runs Claude Code for you, a playbook you can press Play on. Each step has a prompt, a kind such as a regular step or an approval gate, and a connector that runs the next step sequentially, in parallel or conditionally. Recipes live on disk and are built visually. Workflows (visual automation canvas, in development) An in-development n8n/Zapier-style automation builder: a workflow is a trigger plus a graph of steps, edited on a canvas beside the projects sidebar, with a step picker, a settings drawer, live validation, Build with AI that turns a sentence into a graph, a split view that watches an agent build it, retries and on-failure paths, and a wait-for-me approval step. Zapier Integration Connects Claude sessions to thousands of apps through Zapier's remote MCP server. Once it is enabled, every spawned session gets Zapier tools injected automatically, so an agent can trigger Zaps, read data from connected apps, and manage workflows with no per-session setup. Setup is one server URL pasted into Settings, and the app connections stay inside Zapier. Scheduling Alarm Quick Add A tab inside the Quick Launch floating composer that creates an Omniscio alarm from one natural-language line such as "every weekday at 7:30 AM", covering the phrasings the parser understands, the defaults it fills in, the preview row's status messages, and how to edit the alarm afterwards. Alarms The Alarms feature: natural-language reminders that ring inside Omniscio at a wall-clock time you choose, with snooze and dismiss, recurrence shapes, per-alarm sounds and folders, a missed-fire inbox row, and the CLI routes an agent drives them with. Alarms (part 2) The advanced half of the Alarms page: the snooze cap, the math challenge that gates Dismiss, Test fire previews, fire limits, missed alarms in the inbox, the mobile versus desktop split, the natural-language parse cost cap, and the CLI routes. Create a cron job by asking AI Omniscio's built-in cron scheduler can be driven by an outside AI: describe the schedule in plain English and the AI creates the job, which then waits for your approval before it ever runs. Covers enabling CLI Control, phrasing one-off and limited jobs, the approval step, and how to revoke the AI's access to your machine. Cron job — Session spawn type The Session job type in Omniscio's cron scheduler: instead of running a shell command or a recipe, it starts a brand-new Claude session in a project at the scheduled time, opening with a prompt you wrote, or wakes an existing session on a schedule. Covers setting one up, how each run is numbered, the live-session guard, the five-minute floor, and the heartbeat governor that nudges, cards and finally retires a wake schedule whose rounds keep finding nothing. Cron jobs (scheduling work to run on its own) The Cron Jobs screen: how to create and manage a scheduled job by hand — the three job types (Script, Recipe, Session) and the fields each one takes, the schedule, approval and self-healing switches, the jobs dashboard with its search and grouping, per-job run history, and Run Now. The companion page covers handing a job's description to an AI to create. Also covers fire-time timing (a start can be up to ~2 minutes late under load) and the six alerts it can raise. Cron self-healing How Omniscio reacts when a scheduled job fails for good: it pauses the job and hands the failure to a fresh Claude session pre-loaded with the error and logs, so the agent can fix the cause, verify it, and resume the schedule — plus the three-strike cap that stops it looping. Drip (queue-and-trickle inbox feeder) Drip holds mixed content — notes, links, files, whole folders — in named queues and releases items into your inbox on a cadence you set, so a sixty-link reading list arrives a few at a time instead of all at once. Drip (queue-and-trickle inbox feeder) (part 2) The second half of the Drip page: the internals behind a drip — its tables, services and CLI surface — and what the inbox does with a released item, from previews and thumbnails to right-click actions and editing. Night Shift (a queued plan of work that runs while you sleep) Queue an ordered plan of overnight tasks and Night Shift runs the phases one after another while you are away — waiting for each stage to finish, watching the sessions it started, and noticing when one stalls instead of silently hanging. You come back to a finished run and a report rather than a stopped queue. Nighty Tidy (an overnight audit that only touches its own work) An overnight audit panel that tidies and checks things while you sleep, built around one hard boundary: it owns what it created and leaves the rest of your work alone. Covers that boundary, how to run one, and what it will refuse to touch. Nighty Tidy — how it works and its limits (part 2) Part 2 of the Nighty Tidy page: the machinery under an overnight run, the sessions it tags as its own, how to drive it from an AI operator session, and the boundaries you should know about before trusting it with an unattended night's work. Rename a cron group header How to override the label a group of cron jobs shows in the Cron Jobs view and its sidebar without renaming the project itself — the pencil on the group header, the 80-character limit, what the override does not touch, and how orphaned labels are cleaned up. Run if missed (cron catch-up) A per-job switch, off by default, that makes a cron job catch up once when Omniscio was closed across its scheduled time — instead of quietly skipping the missed fire as it does today. Covers where the toggle lives, why every missed window collapses into a single run, the seven-day limit, and what a catch-up run still respects. Scheduled Messages (compose-to-self, fire-into-inbox) Scheduled Messages lets you write a rich-markdown note to yourself, schedule it to arrive once or on a recurring cadence, and have Omniscio drop it silently into your unified inbox when it fires — a letter to your future self with no approvals and nothing to respond to. Scheduling a recipe How to put a recipe on a calendar: the Schedule Recipe dialog, the four frequencies (Interval, Daily, Weekly, Monthly) and every control each one shows, the quick-select interval presets and the 15-minute floor, the time and day-of-month pickers, the Next run preview, and how to pause, change or remove a schedule afterwards. Recipe runs cost money, so the defaults matter. Dev Pipeline Auto-lander dashboard The Auto-lander dashboard — the Dev Pipeline panel's Auto-lander tab, where you watch the auto-lander's live status, see where each of your branches is and what it is waiting on, read the history of what it has landed or handed back, and pause or resume the lander or a single repository. Base-branch picker How the Base branch picker in the Start session dialog works: choosing which git branch a new session's worktree forks from and merges back into, where the dropdown appears, how the choice is remembered on the session, and the one boundary worth knowing about the auto-lander. Clearing a workspace that sits outside every folder (the one removal Omniscio will not do by itself) How a workspace (git worktree) that sits outside every folder Omniscio manages gets cleared, now that nothing else can touch it: the inbox card that reports it carries a button, clicking the button raises one approval per workspace, and approving an approval removes it — folder gone, branch kept. Covers what the app refuses to remove even then, what the removal still checks, and how an agent reaches the same outcome from the CLI. Dev Pipeline The opt-in skill that runs a whole software-development workflow on one task — worktree, investigate, plan, red-team, build, elegance pass, docs, git-prep — pausing at five approval gates, with each phase optional and configurable. Dev Pipeline (part 2) The opt-in behaviour modes and configuration behind Dev Pipeline — independent-review, the AI Code Review phase, tiered-model execution, per-phase and project-level custom instructions, the always-on standards phases, the auto-approve marker, and how the skill runs self-contained. Dev Pipeline Maintenance The two housekeeping skills that ride with the Dev Pipeline: worktree-cleanup reaps finished worktrees and merge-all-ready lands what is ready. What the unattended daily run does, what the verifier checks, and how to stop it. Dev Pipeline panel The built-in sidebar panel that oversees the dev-pipeline workflow in one place: live runs and the worktrees they hold, per-run timing and cost, the Agent Board, one-click gate approval and safe worktree cleanup. Dev Pipeline panel (part 2) The second half of the Dev Pipeline panel page: the Auto-lander tab, the QA Fleet tab and the Setup tab — the switches, the cleanup jobs and the spend caps — plus how the panel is put together for contributors. Developer Guardrails The machine-wide sibling of Git guardrails: a set of rules written into your global Claude Code settings so every session on the computer — not just the ones Omniscio spawns — is held to them, with escape hatches and a warn-only mode. Gauntlet Loop (build, blind-judge, repeat until a builder passes) A developer panel that runs several builder agents at one goal, has fresh blind critics score every attempt against a concrete bar, and loops until an attempt clears your pass score or the run hits its round ceiling or cost cap. Git guardrails The rules that stop a session doing something destructive to your git history — committing to a protected branch, force-pushing, or throwing away work — with a warn-only mode, per-action approval, and escape hatches for when you do mean it. Git guardrails (part 2) The guard rules git-guardrails gained after launch, each dated and named, from the first September additions through the 2026-09-30 pair — the running record of what the guard blocks today, kept as a continuation of part 1. Hotspot Report (which hand-authored files are too many people editing) Developer infrastructure: the readout that finds the hand-authored files many agents edit at once — the design defect behind most sync conflicts. It ranks by distinct commits in a window, keeps generated files out of the ranking and lists them separately as rule-zero offenders, and names a one-line split proposal per hotspot. Read-only, always exits 0. Markdown merge driver (why two agents can edit one document) Developer infrastructure rather than an app feature: a custom git merge driver that resolves the most common conflict in this repository — two agents independently appending to the same memory document — and falls back to ordinary conflict markers for anything it cannot resolve safely. There is no screen and no setting; it runs inside git. Master Sync (keeping your local master current) Developer infrastructure, not an end-user Omniscio feature: the one in-app job that keeps this machine's local `master` branch current with everyone else's work on GitHub. It runs every 15 minutes by default, needs no setup, and only ever spends a "merge agent" session when a conflict genuinely needs a judgment call — capped at a number you control per day. The same page covers the catch-up road: what to run when you are too far behind for the sync to finish. Master-Debt Auto-Fixer A long-lived shared branch accumulates debt: a test, lint rule or build that has been failing for a while and belongs to no single branch. This watches for failures that already exist on the shared branch, waits until the same one has survived three separate runs, and then gets it fixed — per file, and only when it sticks around. Orphaned Workspace Archive (reclaim the folder, keep the work) A background job that reclaims the disk folder of a workspace whose session has been gone for a day, while keeping its branch and every change on it — uncommitted work is committed first and also saved to a recovery copy. Covers what it refuses to touch, why the first run only reports, and how to get the work back. PR conflict conversation When a pull request you pushed conflicts, the agents involved settle it in one short conversation: the main computer opens it, everyone keeps syncing on GitHub master's copy, the responsible side tests and updates the pull request, and every computer adopts the published result. Your computers talk over Agents across your computers, other developers over their agent lane; it exchanges records, not prose, and never removes unpublished work. PR Janitor The nightly job that closes GitHub pull requests whose work is already wholly on origin/master, investigates everything it leaves flagged, starts landing sessions for whatever still holds unlanded value, and comments on every PR it judged — plus the operator seat, run ledger, dry run and pause switch inside the PR Merge Queue panel. PR Merge Queue Omniscio's PR Merge Queue: an in-app view listing every open pull request across your configured repos, sorted into lanes by risk, with one-click sessions that prepare a merge in an isolated worktree. Covers the master flag, the toolbar and sidebar mounts, the lanes, background auto-refresh, Inbox nudges, archiving, batch merge and the current limits. PR Merge Queue (part 2) The deeper half of the PR Merge Queue page: what the feature persists, the in-flight guard that stops a PR being merged twice, custom and per-repo merge instructions, editing the merge prompt, the pure-logic triage modules, the lane decision tree, the IPC channels, the auto-merge daemon and AI summaries. PR Visual Evidence Before/after screenshots and short screen recordings of a UI change, captured at the end of a coding-agent session and attached to your private session by default — and to the public GitHub PR only when the repo opts in a second time. Covers the per-repo switches, both capture paths, and the safety rules the platform enforces. Publishing the websites How Omniscio's public websites — the documentation site, the homepage, the admin console, the team-chat app, the web demo and two short-link redirects — actually get published: from the developer's own computer after every push, by one publisher, now that the GitHub jobs that used to do it have been stopped and the hosting one deleted. Repo Foundations Repo Foundations gives any repository the foundations that make it productive for AI agents: a documentation system an agent can navigate from a cold start, a Claude-rules baseline that keeps its work safe and consistent, and a cross-tool setup so those foundations reach every AI tool you use. It inspects the repo, previews every file, and writes nothing until you approve. Restart offer (when a sync pulls new code under a running dev app) Running Omniscio from source (npm run dev), the app builds whatever is on disk when it starts. If you also run the 2-hourly Sync Master job, that job then rewrites those same files underneath the running app — it pulls the latest master into your checkout. Nothing about that reaches the window you are looking at. Review changes (diff viewer) Omniscio's built-in git diff viewer: browse every changed file in a project or a single session, read the diff with review overlays, and accept or reject changes hunk-by-hunk — plus committing and pushing from the same panel. Review session titles (how an AI reviewer session is named) How a session started to act as an AI reviewer is titled: one fixed shape, `[<Kind> Review] <AI>: <what it reviews>`, where the app writes the review type and the AI's name and the agent that asked for the review writes the subject. Covers what each part means, where the titles appear, and what an agent sends to get one. Session auto-lander queue-status notice A small in-thread notice at the foot of a conversation, telling you where a finished session's branch sits in the auto-lander's merge queue: next up to land, queued at number N of M, landing automatically with a stated reason, or amber when it genuinely needs you. Purely informational, with no buttons. The drift chip in the sidebar footer (and its warning levels) The small pill at the bottom of the sessions sidebar that reports drift on three axes at once — the code on disk against the code this window is running, this checkout against its remote, and how many finished branches are still waiting to land — plus a fourth reading of how many worktrees are in use against how many are parked. What each arrow, circle and number means, and how to set your own warning levels. The paved road (git work on a worktree, over the CLI) The "paved road": the CLI routes an agent uses to do git work on its own worktree — commit the files it changed, sync onto the latest shared branch, stamp the branch ready to merge, hold and release a gate, hand a worktree over or take one back, reclaim a stranded one, and repair a broken dependency folder. Explains why the raw git commands are the wrong tool for each job. Unfinished work card (branches nobody is coming back for) RETIRED 2026-09-30 — the drain no longer raises this card; the daily maintenance report took its job. It listed branches still holding commits never merged while the session that worked on them was gone, whether their folder was on disk or had been freed, with two buttons: pick the work back up (a paid agent) or let the folder go (the branch is kept). Before retiring it, 134 of ~144 rows offered only the paid-spawn button, and it re-raised 26 times. The code and the ADOPT/DROP routes stay. Verdict panel The Verdict panel is the one screen that answers whether the verdict road is healthy right now and, if it is not, exactly where the line is backed up: one status sentence, six headline stats each judged against a normal band, the road drawn as eight stations with its bottleneck marked, and one-click drill-downs into every station, the fleet, this machine and the broker machinery. Read-only, in development, also readable by agents. Web-App Verification (Obscura) When an agent builds you a web app, this feature lets it functionally verify that the app actually works before it says it is done, using Obscura, a lightweight Rust headless browser, instead of installing full Chrome. The agent gets one tool that checks the page loads, elements exist, forms accept input, and no console errors appear. Where checks run (the testing setting) One per-machine setting decides where heavy checks (tests, type checks, lint) run: cloud first (this computer is a small, capped backup while the cloud is down), cloud only, or this computer. It is on the Dev Pipeline panel's Setup tab and at the terminal as npm run testing; a person changes it, an agent can only ask on a card. Where's My Work A read-only screen that answers "what actually happened to my changes?" — it reads your project's copy of the repository and groups every branch into plain-language buckets (landed, waiting on review, turned down, stranded), so a reader who does not use git can still see whether finished work reached the shared repo or is sitting on this machine unasked. Worktree Cleanup Skill (reap landed worktrees, safely and reversibly) A Claude Code skill that bulk-cleans the git worktrees left behind by finished work, deciding what has landed by content rather than by branch name so it also catches squash, rebase, cherry-pick and replay landings, and saving uncommitted work plus a recovery SHA before it deletes anything. Knowledge & Memory 26 What Omniscio remembers — your vault, its memory, and what it captures for you.
The Vault (KMS) KMS (the Vault) — notes, editor, search and sharing Omniscio's built-in Markdown vault: point it at a folder of notes and edit them in place, with a file tree, search, wiki-links, images and sharing. This page covers getting around the vault and the editor; the long tail — the editor's details, media, the AI tools and the internals — continues on the parts below. KMS (the Vault) — notes, editor, search and sharing (part 2) Part 2 of the KMS page: the editor itself — its toolbar and slash commands, how links, tables, footnotes, lists, headings and code are written and rendered, how the view behaves while you type and save, and what happens to an imported note that is not perfectly formed. KMS (the Vault) — notes, editor, search and sharing (part 3) Part 3 of the KMS page: how the vault looks and how you find a note in it — the colour and theme options, the sidebar you can hide, wiki-link autocomplete as you type, the unified search modal, and the quick-switcher for jumping straight to a note by name. KMS (the Vault) — notes, editor, search and sharing (part 4) Part 4 of the KMS page: images, media and the links that turn into something — how a picture is added, stored and displayed in a note, how attachments are kept from bloating the vault, and what happens when you paste a bare web address into a note. KMS (the Vault) — notes, editor, search and sharing (part 5) Part 5 of the KMS page: what the AI adds to the vault — the note summaries and agent tools, asking a question seeded from a note, keeping the vault loaded so it opens instantly, the standalone window, deep links to a note, and publishing one to a public web address. KMS (the Vault) — notes, editor, search and sharing (part 6) Part 6 of the KMS page: using the vault away from the desktop and under the hood — the mobile experience, what happens after a crash, what is deliberately not built yet, the command surface an agent or script calls, how the vault is put together, and what it promises about your notes. KMS agent tools — expose your Markdown vault to spawned agents An opt-in extension that lets the Claude sessions Omniscio starts read and write your KMS vault through agent tools — searching, fetching, listing and traversing notes, and separately creating, appending and editing them. The read and write halves are gated by two different switches, off by default. KMS callout blocks — Obsidian-compatible admonitions in the editor Styled admonition boxes inside the KMS note editor, written in Obsidian's own callout syntax so a vault written in Obsidian looks the same here and survives a round trip. Twelve types ship, fourteen Obsidian aliases are understood, and a type Omniscio does not know still keeps its original text rather than losing it. KMS knowledge aggregation (periodic AI-curated session summary notes) A background pipeline that turns a period of session activity into curated notes in your vault: it scans recent sessions, picks the notable ones, pulls the key insights out with a cheap model, and writes one summary note per project into an Agent Insights folder. Opt-in and cost-capped. KMS note summaries (AI auto-summary + drift lifecycle) A one-paragraph AI précis stored inside each note's own file, wrapped in a comment block so any Markdown editor round-trips it untouched. Omniscio notices when a note has drifted away from its summary, shows that state on the note itself, and can quietly refresh stale summaries on a timer. KMS Quick Reference Wizard A seven-step walkthrough that introduces the KMS vault to someone opening it for the first time — what the vault is, how notes and organisation work, how to find things, the agent tools, asking your vault a question, and setting up the vault folder. It changes nothing except that final folder, and can be replayed whenever you like. KMS Session Seeds (multi-note briefing + project context notes) Two ways to pre-load KMS notes into a Claude session so the AI starts with real context rather than a blank slate: a multi-note briefing that assembles two to five selected notes into one opening prompt, and project context notes attached to a project so its sessions begin already briefed. KMS Vault Overview (searchable inventory table of every note) A read-only table showing every note in the active vault on one screen — title, tags, summary state and text, estimated token count, size and last-modified — sortable and searchable. It is the bird's-eye complement to the note editor, for auditing a large vault or judging what a note would cost to load. Update Vault (/update-vault skill) A built-in skill that scans your Omniscio sessions and logs what they did into your KMS vault. It extracts PRs shipped, bugs found and fixed, features built, test results, UX improvements, costs, and blockers, then updates the matching vault note for each session's feature or project. It runs on the current session or as a batch scan, and never modifies, archives, or deletes them. Memory & Search Global memory An agent-written memory library the agents can browse, search and add to — modelled on Omniscio's own docs system but live: a budget-aware tree a session can load at the size it wants, memories an agent writes from its own work, and comprehension search. Global Search (Ctrl+K across everything) One keyboard-first palette that searches every message channel in Omniscio at once — conversations, SMS, Slack, Telegram, feeds, webhooks, session titles and project metadata — with quoted phrases, exclusions and deep links into any result. MemPalace (cross-session memory for your sessions) A cross-session memory system that gives every Claude Code session a set of MCP tools for remembering what you have done before: your transcripts are mined in the background and organised hierarchically into wings and rooms an agent can search. Off by default; turning it on wires the tools into newly spawned sessions. Semantic Search (AI-powered meaning-based search) Semantic Search lets Global Search find sessions by meaning rather than keywords, so a query like how do I speed up my deploy can surface a session titled Fixing CI pipeline latency. It runs on a local embedding model entirely on your machine, ships on by default, and blends invisibly into Ctrl+K results. Session Search (sidebar virtual project) Session Search is a permanent sidebar entry: a chat with Claude where you describe a past session in plain English and Claude does the searching, then replies with clickable links straight into the session it found. It is backed by a local embedding index of session titles and first messages, so it can match by meaning rather than keywords. Ctrl+K remains the separate, instant keyword palette. Capture & Import ContextDock (Native) — local-first library (in development) The native ContextDock: a local-first rebuild of the document library that reads Omniscio's own store rather than a remote service, with panes for documents, bundles, lists, tags and assembling a context payload. Covers adding documents and files of any type, editing bundles and lists, the one-time cloud import, and what stays deferred. ContextDock Integration (link bundles and lists to project knowledge) ContextDock is a separate web app where you collect markdown notes into bundles and lists; linking one into an Omniscio project takes a snapshot of it, and every new session in that project starts already knowing the content. This page covers turning the integration on, linking and refreshing what you link, and managing the linked rows. ContextDock Integration (link bundles and lists to project knowledge) (part 2) Part 2 of the ContextDock integration page: what the integration deliberately does not do, the full error-and-recovery map for every failure code, the size caps that decide how much linked content can reach a session's first message, and what to gather before filing a bug. ContextDock Integration (link bundles and lists to project knowledge) (part 3) Part 3 of the ContextDock integration page: where a linked snapshot is written on disk, the sentinel header that marks it as Omniscio-managed, how the picker cache is kept warm, how a snapshot reaches a session's first message, and the full agent-facing IPC surface. KMS Obsidian Import Wizard A guided wizard that copies an Obsidian vault into the Omniscio KMS vault — notes, folder structure, supported images, wiki-links and callouts — while never modifying or deleting the originals. Desktop only. Plugins, themes, canvases, non-image attachments and very large files are deliberately left behind and listed as skipped. KMS Web Clipper (clip web pages and screenshots to the Vault) Save a web page by URL, or the text visible in a screenshot, straight into your KMS vault as a clean Markdown note with its source, timestamp and tags recorded. Pages are fetched and extracted for you; screenshots go through an AI read that costs a small amount. KMS must be enabled. Summarize Video Skill (smrzz to KMS capture) A Claude Code skill (/summarize-video) that takes a video, podcast, or article URL, sends it to the smrzz plugin for AI summarization, creates a single KMS note with the result, and offers a brief follow-up interview for additional extraction (to-dos to Mission Control, separate concept notes, etc.). The default path is zero-friction: drop a link, get a note in Summaries/ in your KMS vault. AI Features 60 The AI surfaces beyond a plain session — coaching, writing help, voice and media.
Assistants & Coaching AI Coaching sidebar (interview prompts and artifacts) AI Coaching is a virtual project in Omniscio where you sit down with Claude for a structured interview and walk away with a saved artifact — how to turn it on, choose a coaching style, work the Dashboard, and what happens to your sessions and bookmarks along the way. AI Coaching sidebar (interview prompts and artifacts) (part 2) Part 2 of the AI Coaching sidebar page: how an interview's artifact is saved, edited, versioned, restored and deleted, how pinned context and detail levels feed artifacts into sessions, and how imported work and the KMS vault mirror fit in. AI Coaching sidebar (interview prompts and artifacts) (part 3) Part 3 of the AI Coaching sidebar page: the free-form coaching sessions, the compaction summaries and durable memory facts behind the coach, the synthesized Core Profile, the Life Inventory surfaces, and the telemetry that proves the memory system works. AI Coaching sidebar (interview prompts and artifacts) (part 4) Part 4 of the AI Coaching sidebar page: the CLI control-server surface for AI Coaching, the developer-only interview prompt editor, and the implementation pointers for agents working on the feature. AI Coaching — FAQ Short answers to the questions people ask most about AI Coaching: whether you need an API key, how artifact detail levels work, quick check-ins, mobile use, importing and exporting, deleting your data, and the 20,000-character limit. AI Coaching — Feature Guide A walkthrough of every part of the AI Coaching space: the dashboard of interview topics, semantic search, your sessions, the artifact viewer with version history and detail levels, bookmarks, and the Core Profile. It also covers coaching style, archived sessions, and how artifacts reach your regular sessions. AI Coaching — Getting Started AI Coaching is a guided conversation space inside Omniscio where you sit down with Claude for a structured interview about a topic that matters to you, and walk away with a written summary you can keep. This page goes from turning it on to your first interview. AI Council launcher (one-click start a council from your default panel) The AI Council launcher in the Prompt Tools sidebar group is a one-click action row that opens the AI Council and immediately starts a new council from your default panel, falling back to the guided setup wizard on first use, plus where the optional Councils section shows in a hub's session list. Ask about this page (floating screen-aware helper) A floating, screen-aware chat helper: how to reveal it and open it, the first-use consent gate, exactly what gets captured from your screen and what never does, its private history, and why its chats stay out of the sidebar and inbox. Ask Omniscio — in-app help agent The sidebar help agent that answers questions about Omniscio from the bundled library and your live state: which model it runs on, what it knows and can do, the question log, the privacy model, and its v1 limits. Chief of Staff — the guided assistant for non-technical users Chief of Staff turns AMC into a single friendly chatbox. Instead of learning projects, sessions, panels, and settings, you turn on one toggle — or just click the Chief of Staff item in the sidebar — and talk to it in plain language — "remind me to call the vet at 4", "save this as a note", "what needs my attention right now?" — and it does the work for you behind the scenes. It's built for people who want the power of AMC without the cockpit. Get Help (Helpdesk) Omniscio's in-app help assistant: ask anything and get an answer drawn from its own documentation with clickable source chips, escalate to a human when that is not enough, and follow the conversation from question to resolution. Get Help (Helpdesk) (part 2) The second half of the Get Help page: the operator queue and developer console the team answers from — status tabs, drafts, routing and aging rules — plus the boot-time listener, the alerting watcher and the key files. Help Desk autopilot The Help Desk autopilot: a help-desk plan that only fixes a confirmed bug is approved without waiting for a person, and the reporter is told in plain English when their fix lands and again when it ships. What the autofix-or-hold rubric is, what always waits for you, and the one switch. Help Desk email tickets — a plain-English support email becomes a tracked ticket Someone emails your support address in plain words and it becomes a real Help Desk ticket: it appears in the operator's incoming queue, advances a lifecycle stage on its own, and the reply goes back out as the support address on the reporter's own thread. Jev — a decision model your agents can call Jev is a decision model your agents can call when a task needs a choice rather than prose: you hand it a situation and typed questions, and it answers each one with the confidence it assigns. This page covers what it is for, how an agent runs it, what it costs, and how to read a confidence before you act on it. Life Inventory (AI Coaching self-assessment) A structured self-assessment inside AI Coaching: instead of talking everything out, you rate yourself one to five across two ready-made inventories — 942 symptom items and 1,246 strength items, grouped into categories. Your ratings are saved locally as a take you can repeat later to see what has changed. Local Chat (chat with a local model via Ollama) A chat area for talking to a model that runs on your own computer — free, private and offline. It is deliberately not a coding session: no tools, no file access, nothing that edits your project. Replies stream in, conversations are saved and renameable, and nothing you type leaves the machine. Multi-Model Council (ask several models at once and get one answer) Put a question to several models in parallel and get back a single synthesized verdict, with the individual answers kept alongside it. Covers creating a council, the guided setup, the two ways it can be run, and how the panel is assembled. Multi-Model Council — asking, verdicts and cost (part 2) Part 2 of the Multi-Model Council page: putting a question to the council from a session, what a verdict actually shows you, the actions available on each turn, debate mode, attaching files, how the conversation remembers itself, what a run costs, and where the code lives. Report conversations — replies to your reports, inside the app Every report sent from the app's Feedback button also opens a conversation with the support team, so support can answer inside the app and the person can write back — even without Get Help. For a report of any kind from somebody other than the operator, that conversation is the whole answer: no investigation, one acknowledgement, and a draft reply waiting for support. Covers where replies show up ("Your reports"), what the conversation carries, and how it is switched off. Starting a session from a Help Desk ticket Two ways a Help Desk ticket becomes a Claude Code session: you start one from the open ticket, and a newly arrived ticket starts a read-only investigation by itself. Neither one writes to the ticket or reaches the customer. Super Prompt Creator Super Prompt Creator is a one-click way to design a brand-new super-prompt. It launches the builtin superprompt-creator super-prompt (the "PromptArchitect" persona) in a fresh session: it interviews you about what you want the prompt to do, then generates a complete, ready-to-use super-prompt you can save and reuse. Support Chat In-app chat for Omniscio users to contact support directly from the toolbar. User mode (default): A chat icon in the toolbar opens a floating panel (bottom-right). Users enter a display name on first use, then send messages that reach the operator's inbox. Operator mode (supportChatOperator: true in settings): The operator sees all conversations as inbox items and can reply from the same floating panel or the inbox. Unclaimed-check capability (sign-in-gated US unclaimed-property lookup) A free, sign-in-gated capability that lets an Omniscio user's agents check whether a person is owed US unclaimed property by name. It covers California and New York today, runs against a publicly published state-records service, and every result links to the official state site to claim for free, never a paid finder. Use Skills (browse, install, edit) A Skill is a small markdown file that teaches Claude Code how to do a specific thing, such as shipping a package or writing a postmortem. Omniscio ships an in-app Skills browser for installing skills from a curated catalog, editing your own, and keeping them current. Installed skills activate automatically when their triggers match. Use Super Prompts (reusable prompt library) Super Prompts is a built-in library of 190 reusable first-message prompts for the Claude virtual project: search or browse them, pin favourites, duplicate and edit them, author your own, hide the ones you do not want, and launch a session already briefed with a well-crafted prompt. Voice & Speech FlowVoice (OS-wide dictation + history) Omniscio's OS-wide dictation: hold a hotkey, speak, and the cleaned transcript is typed at the cursor in whatever app is focused — a browser, an editor, Slack — not just Omniscio's own message box. Covers the two hotkey slots, the hosted and Gemini speech engines and what each costs, the floating HUD and composer mic button, where to configure it, and the promise that a dictation is delivered whole or its audio is kept. Microphone calibration for voice barge-in (tune sensitivity per mic) Microphone calibration is a short guided test that measures your mic and your room once and tunes voice barge-in sensitivity to them, so a fan or a keyboard no longer cuts the AI off and a quiet speaker still registers. It sits behind an off-by-default Lab flag inside the Voice Report Back settings card, is stored per microphone, and saves nothing when you skip it or when it cannot measure cleanly. Reply by Voice (talk back to a spoken recap) The reply sibling of Spoken Narration: a mic button beside a message's narration play button opens a window where you talk your answer instead of typing it, reviewing your words before they send. In development and off by default, and it needs Voice Input turned on to capture anything. Spoken Narration Spoken Narration is the audio sibling of Plain Speak. Where Plain Speak shows a rewritten glance-card on screen, Spoken Narration plays a short, friendly spoken recap of each completed agent reply — in the agent's own words ("here's what I did …"), not the raw message text and not the general read-aloud summary. It reuses the existing TTS pipeline (Fish Audio / Grok / ElevenLabs). Every engine can produce it (2026-07-31). Voice "About to Act" Highlight (voiceActHighlightEnabled) When you use voice conversation and ask the assistant to do something that needs your confirmation, the assistant speaks a yes or no prompt before acting and this feature also draws a spotlight on the on-screen thread it is about to act on. The highlight is purely visual, never performs the action, and is off by default. Voice Attachment (Omniscio Voice vs Session Voice indicator + pin) Voice can target the app-level assistant or one session, and this feature makes that target visible with a small chip and lets you pin voice to a session so it stays there while you browse elsewhere. It is off by default, and it adds no new AI call and no new IPC: the chip and the route receipt derive from state that already exists. Voice History A viewable, persisted history of your voice conversations, kept in the Omniscio sidebar group so you can revisit what was said. It holds both the general hands-free assistant and the per-session voice conversations in one list, distinguished by a badge, and deleting a conversation erases its spoken transcript immediately and permanently. Voice L3 - Ask About a Thread (session:converse) Voice L3 lets you ask hands-free questions about a named agent session without opening it or interrupting it: say the wake word or press Alt+V, name the thread, and Omniscio either speaks its Omni briefing or answers a specific question using only that thread's transcript. Covers how names resolve, the bounded spoken follow-up loop, the offer to hand an unanswerable question to the live agent, and the daily cost cap. Voice L4 - Ask the Live Agent (session:ask-agent) Voice L4 lets you ask the live coding agent in a thread a question hands-free, and tells you when it has answered. It holds the question until that agent goes idle, then speaks a short cue, shows a desktop notification, and reads the full answer aloud on request. It works on running threads only and never starts a new agent. Voice Report Back (TTS read-back of voice-command answers) Voice Report Back speaks the prose answer to a voice command aloud through whichever TTS provider you have selected, so the press-talk-listen loop never needs the screen. Off by default, it speaks only LLM-layer replies, clips long answers behind a Speak more button, and stops at a daily cost cap. Also lists what deliberately stays silent and what this first slice cannot do yet. Voice, TTS & Wake Word (hands-free Omniscio) Omniscio ships a three-piece voice stack so you can run it without touching keyboard or mouse. Text-to-speech reads agent messages aloud on demand or automatically, voice dictation delivers your speech into the active session's composer, and voice commands turn spoken sentences into actions. A wake word listens on device. Voice, TTS & Wake Word (hands-free Omniscio) (part 2) Part 2 of the Voice, TTS & Wake Word page: the engines behind the voice stack. It covers the wake-word engines and the retired Picovoice option, the speech-to-text providers with their keys and costs, the Omni Session Briefing, spoken output that follows the content's language, third-party dictation and spoken thinking fillers. Voiceprint Studio Voiceprint Studio builds and sharpens an email style guide, the document you paste into any AI so it drafts email in your own writing voice. It grounds each draft in your real sent emails, can self-train against a held-out test set, and turns each miss into an approved rule, across four tabs. Still in development, and it never sends email itself. Media & Capture File Conversion Convert a file from one format to another entirely on your own machine — PDF and Office documents to Markdown, images, audio and video, and data between CSV, JSON and YAML — from the app, from a download link, or by asking an agent. Image Studio (embedded website) Omniscio can embed the live JLS Trading image-generation website as a first-class sidebar panel, so you use the real site without leaving the app. You sign in to it with your own account, the sign-in persists across restarts, and it is off until you turn it on in Settings. Image Studio (native panel — generate, compare, library, stats, templates, post-processing) An image studio built into Omniscio: write a prompt and generate, compare several models on the same prompt, browse and search your whole generation history, watch what you have spent, save reusable templates, collect images into named galleries you order by hand, and post-process finished images. All in one panel, no browser. Image Studio MCP (image generation for Claude CLI sessions) A built-in tool server that gives any Claude session twelve tools for working with images: generating from a prompt, comparing models, browsing your history, reusing prompt templates, and publishing an image as a shareable link. It replaced an earlier tool of the same name, so old prompts still work. Record for Agent A mode on top of the recorder: record your screen while you talk, mark the moments that matter by hotkey, typed note, voice or a drawn box, then hit Send to agent. The agent reviews what you showed it, shows cropped screenshots to confirm it understood, and you talk it into a task list. It never files anything — the conversation is the whole output. Screen Recorder (record, edit and share your screen) Screen Recorder is the video half of Omniscio's built-in screen capture — a built-in Loom alternative. Capture your screen with optional microphone and webcam, review the result, trim and rename it in a built-in editor, then share a public link or publish to Vimeo. This page covers the toggle, the tabbed settings, the hotkeys, the editor and the library. Screen Recorder (record, edit and share your screen) (part 3) The parts of screen recording that sit beyond the basic take: presets and templates with Quick Capture, auto-titling, and the features deliberately deferred — the continuation of part 1. Screen Recorder (record, edit and share your screen) (part 4) The library and storage half of Screen Recorder — the panel's two columns, the library view, storage and retention, saved edits and sharing privacy, and the six global hotkeys — the continuation of part 1. Screen Recorder — part 2 (projects, the data model and the agent surface) The second half of the Screen Recorder page: multi-asset Projects, the data model a recording is stored in, importing a video you did not record, letting an agent read frames, and where the recorder's own channels live. Part 1 covers recording, the editor and sharing. Screenshot / Snip Tool The Snip Tool is the image half of Omniscio's built-in screen capture: trigger a snip and a region selector appears; when you finish selecting, Omniscio saves the PNG to the Screenshots library, copies it to the clipboard, and fires a shutter sound and toast per your notification settings. Shared recording rich previews + embeddable player When you share a screen recording by link, the link now shows a rich preview — a thumbnail, the recording's title, and a short description — when you paste it into Slack, Notion, Gmail, and other apps that unfurl links, instead of a bare URL. The recording can also be embedded as a small inline player on another site (a Notion embed block, a blog, a Twitter/X player card) via a dedicated /embed link. Teach Recorder A built-in, off-by-default feature for recording what you do on Windows — clicks, keystrokes, and screen state — and turning that recording into a reusable AI skill that agents can replay on your behalf. The video player The one video player every Omniscio video plays in — in the app, on a shared link, on the page other websites embed, and for films drawn live by code. Covers its controls, keyboard shortcuts, full screen, phones, and what happens to links shared before it existed. Video playback speed (remembered between plays) Every video in Omniscio plays in the one video player, and its speed control sits in the control bar along the bottom of the video. Inside the app the speed you pick is remembered for the next video and after a restart — including a share opened in the app's share viewer; a shared link opened in a browser has the same control but starts at normal speed. Whole-Conversation Capture Whole-Conversation Capture takes a screenshot of a session's entire chat from the session's ⋯ menu. It shows its progress from the click, can be cancelled, and lands like any other screenshot — on your clipboard and in the Screenshots library, with the usual notice. One picture when the conversation fits; several, in order, when it is too long. Covers where the command lives, what you get, and the limits worth knowing. Bake-Off (fan-out) Run the same prompt against several models at once, side by side, and keep the answer that wins — how to open a Bake-Off, save a setup as a preset, let contestants see each other's work, and what stops it running away with your spend. Browser Logins (log in once, every agent reuses it) Log into a site once by hand and every agent Omniscio runs can use that login afterwards — many at the same time, each on its own private copy. Covers the panel and how to capture a login, why passkeys cannot be used in the capture window, how agents pick a saved login up, the three freshness checks, and what is stored where. Embedded Browser Omniscio's built-in web browser, opened from the sidebar as a panel of its own — what the toolbar does, how cookies and logins persist between visits, and the limits that make it a quick-look surface rather than a replacement for your browser. My Real Chrome v2 (publishable — scripting + native messaging, no debugger) The clean-room, Web-Store-approvable rebuild of Omniscio's Chrome bridge, driving your real signed-in Chrome without the debugger permission — how to enable it, what it deliberately cannot do, and why driving it does not leave you anywhere else — the one switch a drive makes is forced by Chrome, lasts for the length of the command, and puts your own tab back. My Real Chrome — Full (RETIRED — see My Real Chrome v2) Retired. My Real Chrome let an Omniscio agent drive your real, already-logged-in Chrome through a small companion Chrome extension, reaching the signed-in sessions the built-in browser cannot. It needed a high-risk browser permission the Chrome Web Store will not approve, so it is now hidden for everyone and replaced by My Real Chrome v2. Kept for reference only. Integrations 100 Connecting the other products you use, including the AI providers that run sessions.
Email & Calendar Agent Email Give your agent its own real email address on a shared domain — claiming a name, deciding who may start a session by writing in, holding blocked senders for review, how mail flows in and out, and the quiet arrival notice it raises in your inbox. Archive Insights AI Digest (opt-in AI cleanup suggestions for the senders you archive) SuperMail already tracks your Archive Insights — which senders and domains you archive, whether you opened them first, and how long you looked. The AI Digest turns that into concrete cleanup suggestions: an AI reviews the senders you archive (especially the ones you never open) and recommends what to do about each — set up a filter, unsubscribe, or mute — and every suggestion has a one-click Create filter button that opens the filter editor pre-filled. Calendar Quick Add Calendar Quick Add is the Calendar tab inside the Quick Launch box: type one natural-language line describing an event and it creates it on Google Calendar, honouring typed time zones and offering a repeat fallback and a calendar picker. Calendar Quick Add (part 2) Part 2 of the Calendar Quick Add page: why the parser is regex-first with the AI as a rare fallback, why the preview shows more than one row only when the input is genuinely ambiguous, and the repo-aware code map underneath the tab. Email Cleanup The Gmail Clean Up tool: it works out who matters to you, then walks you through archiving, unsubscribing and deleting in bulk — with the safety guarantees that keep a bulk action from being irreversible. Email inbound prescreen — editable prompt The safety classifier every inbound email passes through before an agent session sees the body — what you can and cannot edit in its prompt, what the default rules cover, and how a decision is reached and reported. Email Summarizer (forwarded mail + Gmail labels) Omniscio summarizes incoming email and sends the summary back to you, over two source paths that share one rule table, one daily cost cap and one prompt: forwarded mail to a dedicated inbox, and a Gmail label on threads you tag. Forward email transport The forward-email automation action sends one email per recipient when its rule fires, and it can do that through Gmail or through AgentMail. Which transport to pick, what is checked before sending, and what happens when a send fails. Gmail integration (triage inbox inside Omniscio) Gmail inside Omniscio as a virtual project: triage and reply without leaving the app, AI-suggested replies on every message, keyboard-driven triage, and an inbound-security pipeline that screens for prompt injection before an agent sees the body. Google Integrations (Calendar, Drive, Sheets) Calendar, Drive and Sheets behind one shared Connect Google consent — and all three are agent-facing, so you ask in plain English and Claude calls the right tool. What the Calendar panel shows and how syncing and notifications behave. Google Meet Your Google Meet meeting history inside Omniscio: conference records, participant rosters, cloud recordings and speaker-attributed transcripts, synced in the background and browsable from a panel of their own. Meeting Rooms (one-click speed-dial for your meeting links) A speed-dial for video conferencing links: store your recurring meeting URLs and people's personal room links once, then launch any of them with one click or a keyboard shortcut. Works across Zoom, Google Meet, Teams, Skype, Webex and any other meeting URL, with the platform detected from the link itself. Meetings (record a meeting and get notes out of it) Turn a meeting into a written record: Omniscio records it, transcribes it, and gives you notes you can read, search and act on — with the people and companies involved tracked alongside. Covers turning it on, choosing which transcription engine runs, what each of the three screens is for, and what is sent where. Quick Email (Control Space → Email tab) Quick Email is the Email tab inside Quick Launch (Ctrl+Space): add up to ten recipients, type a body, and press Ctrl/Cmd+Enter to send it through your connected Gmail after a short undoable delay. It covers opening the tab, Cc and Bcc, attachments, send-later, saved drafts, per-recipient hotkeys, the global Ctrl+Z undo, and every Quick Email setting that changes any of it. Quick Email (Control Space → Email tab) (part 2) The second half of the Quick Email page: the machinery behind it — the tab component and its fixed window, the main-side send pipeline and its undo timer, the durable send-later path, contact suggestion ranking and live typeahead, the global Ctrl+Z shortcut bridge, the settings schema and the key files. Recurring Special Events Add birthdays, anniversaries and holidays once, and Omniscio creates the Google Calendar entries and lead-time reminders for them every year, automatically. It is a third tab inside the Google Calendar panel that replaces the old reminder spreadsheet, and it can do what a spreadsheet cannot — like ten days before a holiday that moves each year. Set up Email Inbound (receive email into Claude sessions) Omniscio can receive email at a dedicated inbox, run each message through a safety classifier, and either start a new Claude session for it or continue an existing email thread — with the agent's replies going back out as email on the same thread. Setting it up needs an AgentMail inbox, an API key stored in your OS credential store, and the inbox address pasted into Omniscio. Supermail A Superhuman-style email client - folders, compose, snooze, follow-ups, splits, flows, and filters on top of Gmail or IMAP - brought into Omniscio as a native sidebar feature and shipped from the same in-repo source as the standalone web client at mail.jls.dev. It is a React port of the vendored Supermail UI, not a webview or an embedded plugin. Supermail (part 2) Three refinements to the inbox list (the first two apply to the dense single-line rows shown when the reading pane is off): Soft edge fade instead of a trailing "…". When a row's text is too long for one line it now fades out softly at the right edge rather than ending in a "…". Supermail (part 3) The authoritative gap analysis + per-feature status lives in src/plugins/supermail/SUPERHUMAN_PARITY.md. As of 2026-06-23 the client is at near-parity; the non-AI gaps are closed end-to-end: Undo send is a true deferred send — the composer sends with undoWindowSeconds, the backend holds it on a pendingsends row and fires the real sendMessage when the window elapses; Undo cancels it server-side (DELETE /messages/pending/:id). Supermail AI Filtering AI Filtering lets AI read each genuinely-new email in Supermail and triage it against a plain-language instruction you write — the same "AI reads it and decides what to do" idea the Inbox Pilot already applies to your Claude sessions, now applied to your actual email. Supermail AI Summary ("Catch me up") Catch me up is an optional AI summary in Supermail. Open a thread, and it reads the conversation and tells you, in two parts, what is going on: The thread so far — a short summary of the earlier messages in the conversation. The latest email — what the most recent message says and any action it asks of you. Supermail Archive Insights Archive Insights tracks the emails you archive in Supermail — whether you opened them first, and how long you spent — and surfaces the senders and subject lines you keep archiving (especially the ones you never open) — the mail you are effectively filtering by hand, and could turn into a real auto-archive Filter. It is the passive, behavioural counterpart to Supermail's rule-based Filters and its AI-powered AI Filtering: those act on incoming mail; Supermail Assistants (inbox-wide AI launchers) The Assistants strip is a row of one-click AI launchers at the top of the Supermail Sessions tab. Each button starts a purpose-built AI session on your live mailbox that opens with an instant pre-written greeting. It's the AI-session sibling of Supermail's other AI features (AI Filtering triages new mail; Catch me up summarizes a thread) — where those are one-shot helpers, an Assistant is a full working conversation that reads your inbox and helps you act on it. Supermail Contact Groups Contact groups let you name a reusable set of people once, then drop the whole group into a Supermail recipient field (To / Cc / Bcc) instead of typing everyone's address every time. Pick a group and each person is added as an individual, removable recipient. There are TWO sources, side by side: Your Google contact groups — the groups you already keep in Google Contacts, shown READ-ONLY. You manage them in Google Contacts; Supermail just lets you use them. Supermail email rendering — instant open + dark mode How Supermail (the Superhuman-style email client built into Omniscio) turns a message into pixels: opening an email is near-instant, and in dark mode the emails themselves render dark instead of blasting a white rectangle at you. Supermail Video Summaries (Loom & Vimeo) Video summaries are an optional AI helper in Supermail. When an incoming message contains a Loom or Vimeo video link, it reads the video for you and shows a short, AI-written summary card inside the message — so you can tell what a video is about before deciding whether to watch it. It is in development and off by default — hidden until you turn it on in Settings → Features (the supermailVideoSummariesEnabled Lab toggle). When off, nothing runs and no card appears. Supermail — Inbox Emails & Row Density The Mail and Sessions tab row that previously occupied ~160 px of sidebar space in the Mail view has been moved into a compact header above the email list. The previously-empty sidebar column is gone, giving the email list more horizontal room. Users who have the folder nav pinned ("Show sidebar" ON in Supermail's own Settings → General → Personalization) still see their folder list in that column — the change only eliminates the dead space when the nav is in overlay mode (the default). Supermail — Reading View: Focus Mode & Auto-Hiding Toolbar When you open an email in Supermail, the reading view keeps its chrome minimal so the message itself is the focus: The top action toolbar (Reply, Filter, Focus, Pin) and the bottom action bar (done / snooze / label / trash …) are hidden by default and peek back when you move your mouse to them (or tab into them with the keyboard). A Pin toggle locks both bars open if you'd rather keep them visible all the time. Supermail — Send Behaviors, Reply Focus & Rebindable Shortcuts How the Supermail composer behaves when you reply and send. All of this lives in the native Supermail plugin (supermail.md); the invariants are locked by .claude/memory/contracts/supermail-send-shortcuts-contract.md. Supermail — Sidebar Labels, Settings & Overflow Menu A 2026-08-02 rework of the Supermail plugin's own left-nav sidebar (src/plugins/supermail/ui/), so labels are ordered and tucked out of the way, settings live in a real menu button, and the odd three-line hamburger becomes the standard three-dot (⋯) menu. Zoom API integration (meetings + cloud-recording metadata) Zoom API integration connects Omniscio to your real Zoom account over OAuth so you can browse your meetings and cloud-recording metadata without leaving the app. It is read-only and still in development behind a Lab flag, syncing on a configurable interval. Covers connecting, what syncs, what the panel shows, the local tables and IPC channels, and the limits of this pass. Developer Tools AutoHotkey integration (edit, reload, index .ahk scripts) The AutoHotkey integration — editing .ahk scripts in a folder-backed project with a full Monaco editor, reloading AHK automatically on every save through a signal file, generating a three-document AI script index, the no-code Manager, and the plugin port. Chrome Extension Dev (Autopilot) Chrome Extension Dev (Autopilot) lets an agent build and test a Chrome extension itself by driving a real Chrome over the DevTools Protocol — loading the unpacked extension, clicking the popup, reading the console, reloading after edits, and reporting back with screenshots. Covers the setup, the per-session auto-equip, the Extensions hub, and privacy. Codebase Stats (project health dashboard) A multi-tab report on one real project: its size, health, git activity, dependency risk, AI spend and test coverage, gathered from the filesystem, git, npm and Omniscio's own database into a single modal. Covers opening it, the headline numbers, the per-tab detail, ignoring a noisy finding, and spawning a fix-it session from one. Drive Omniscio Cockpit A developer and QA tool for watching an AI operate Omniscio's own interface — clicking, navigating and exercising features toward a goal you type, with its reasoning narrated beside the app and the ability to take the wheel. GitHub code search A tombstone page: the built-in GitHub code search was removed from Omniscio entirely as part of the in-core GitHub detachment, and the GitHub Integration marketplace plugin is what replaces it. Kept for historical reference only. GitHub Issues A tombstone page: the built-in GitHub Issues panel was removed from Omniscio entirely as part of the in-core GitHub detachment, and the GitHub Integration marketplace plugin is what replaces it. Kept for historical reference only. GitHub Releases A tombstone page: the built-in GitHub Releases view was removed from Omniscio entirely as part of the in-core GitHub detachment, and the GitHub Integration marketplace plugin is what replaces it. Kept for historical reference only. Inbuilt Terminal A real shell terminal inside Omniscio, scoped to a project's folder — open it from the sidebar, keep the last 100 KB of scrollback across a close and reopen, fuzzy-search saved command snippets, and find within the output. In development: it appears only when its Labs toggle is on. Jira board A first-party integration that brings your Jira Cloud board into Omniscio so you can work issues without opening the Jira website: pick a board, see its columns and cards, open an issue's description and comments, and move work. It is gated through the unreleased-feature registry, so it stays hidden until enabled. Linear (in development) A first-party integration that brings your Linear team board into Omniscio so you can work issues without opening the Linear website: connect with a personal API key, see your columns and cards, and open an issue's detail. It is gated through the unreleased-feature registry, so it stays hidden until enabled. MCP Servers (the outside tools your sessions can call) An MCP server is an external tool process a session can talk to, and the MCP Servers view is where you manage your own, install official ones from a catalog, and decide which servers each session actually gets. It also tracks how often each one is really used, so unused servers are easy to spot and remove. Mobbin MCP (real-world design inspiration for recipes) A built-in MCP integration that connects Claude to Mobbin — a searchable library of real-world app screenshots, user flows and UI patterns — so recipes that ask for design inspiration pull from shipped apps rather than from the model's imagination. Off by default; set up from a card in Settings → Sessions. New issue button (GitHub Issues) A tombstone page: the built-in GitHub issue creator was removed from Omniscio entirely as part of the in-core GitHub detachment, and the GitHub Integration marketplace plugin is what replaces it. Kept for historical reference only. Outbound Webhooks (push Omniscio events to your own server) When something happens in Omniscio — an agent finishes, a budget is exceeded, a recipe completes — the app can POST a signed JSON payload to a callback URL you control, so your own systems react in real time instead of polling. Covers configuring one, the events you can subscribe to, and how to verify a delivery really came from here. Pull Requests A cross-repo GitHub pull-request manager: it lists every PR you are involved in, lets you read a PR's description, changed files, checks and review threads, comment, approve, request changes or merge, open a new PR from a branch, and optionally surface PRs waiting on you in the Inbox. Running Apps Running Apps is a built-in sidebar virtual project that lists every dev server Omniscio is running for you: see each live app with its local address, open it inside Omniscio in an embedded browser view, and stop it — without hunting for ports or terminal tabs. Sync config to other AI CLIs (in development) An experimental feature that installs the setup you already manage inside Omniscio — your custom MCP servers, your agent instructions and your actual skills folders — into the other AI coding CLIs on the machine: Codex, Gemini, OpenCode, Cursor, Grok, Antigravity, Pi and Hermes. It targets every tool you have enabled automatically, with no per-tool ticking. Covers what is written where, the safety rules, and the read-only project-context broker pilot. Work & Project Tools Airtable The Airtable integration — turning it on, connecting an account with a personal access token, browsing bases and records from the sidebar, the CLI routes an agent uses, and the inbox and workflow notifications it raises. Asana board An unreleased integration that brings your Asana tasks into Omniscio as a sidebar board: connecting it with a personal access token and a workspace id, the inbox rows for tasks assigned to you, and the workflow trigger and CLI routes it exposes. Canva Integration Connecting your Claude sessions to Canva's design platform: the one-toggle setup, the Canva sign-in a session triggers on first use, the roughly thirty design tools agents gain, and the editing, plan, and rate limits you should know before relying on it. ClickUp ClickUp is a first-party integration that connects to your real ClickUp workspace over the ClickUp API v2 and shows any list as a kanban board in a projects-sidebar row — creating, editing, moving, deleting tasks and comments without a browser. Covers enabling the in-development feature, connecting a token, and what stays read-only. Composio A connector that lets agents act in a thousand-odd outside apps — Gmail, Slack, GitHub, Notion, Linear and the rest — through a service that already holds the logins. Covers connecting an app, how the tooling reaches every new session automatically, what stays on your machine versus the service, and the feature's unreleased status. Dropbox integration (file browser and management) A Dropbox file browser inside Omniscio: connect your account once, then browse, search, upload, download, rename and delete files from a sidebar panel, with the account token kept in your operating system's credential store. Fathom (meeting transcripts via MCP) A built-in MCP integration that gives your sessions read-only access to your Fathom.ai meeting recordings — list recent meetings, search by keyword, read a speaker-labeled transcript and view summaries, without leaving the conversation. Google Docs Export (publish Markdown to Google Docs) Take a Markdown string from anywhere in Omniscio and publish it as a fully-formatted new Google Doc: real headings, tables, lists and code, with the URL handed back as a toast and old copies of the same title tidied away. Google Drive integration (agent-driven file management) Google Drive wired into your sessions instead of a side panel: seven tool-use calls an agent can make to list, search, create, move and share files, over the same single Connect Google consent the other Google integrations use. Google Sheets integration (agent-driven spreadsheets with rich formatting) Omniscio's Google Sheets integration lets you view and edit Google Sheets directly inside Omniscio. The built-in Sheets virtual project renders spreadsheets with their native Google Sheets formatting — background colors, text colors, bold, italic, font size, horizontal alignment, and actual column widths — so the grid inside Omniscio looks like the real spreadsheet, not a plain-text table. Google Workspace MCP (Docs read/edit, Sheets, Drive for Claude CLI) An MCP server that gives Claude CLI sessions direct access to Google Docs, Sheets and Drive — nineteen tools that run as a subprocess alongside the CLI, so the agent reads, searches and edits documents mid-session without Omniscio mediating each call. Granola (meeting notes) An in-development integration that brings Granola meeting notes into Omniscio: a browsable list of recent meetings, a read-only transcript viewer, an inbox row when a new meeting appears, and one click to hand a meeting to a new session as context. Monday.com (connect your real Monday.com account) A first-party integration that connects Omniscio to your real Monday.com account through its Cloud API: view boards, manage items, track assigned work in your inbox, and react to item changes without leaving Omniscio. In development and off by default; connects with a personal API token rather than a sign-in. Notion board (work a Notion database as a board) A first-party integration that brings a Notion database into Omniscio as a kanban board, so you can work its pages without opening the Notion website. In development and hidden by default; connects with a token you paste into the panel, and can also show recently edited pages in your inbox. OneDrive integration (file browser + cloud storage) A sidebar file browser for Microsoft OneDrive, so you can browse, upload, create folders, rename, move, copy, search and share files without leaving Omniscio. Covers connecting it with your own Microsoft sign-in, what you can do once connected, and where the browser appears. Trello A first-party integration that connects to your real Trello account over the Trello REST API and surfaces as a Trello row in the projects sidebar, opening any of your boards as a kanban view. Cards can be read, created, moved, archived and commented on without a browser, and cards assigned to you can land in your inbox. Chat & Comms Discord Import Wizard An in-development, feature-flagged wizard that imports a Discord server's channel history into Team Chat: six steps from the bot token through channel selection and user mapping to a live import with a closing report. Microsoft Teams An experimental in-development integration that pulls your Microsoft Teams online meetings into Omniscio through the Microsoft Graph API. It is read-only, so Omniscio lists, displays and syncs meetings but never creates or modifies them in Teams. Meetings sync every fifteen minutes and can be browsed, opened in full detail, joined, or synced on demand. Slack integration (triage Slack from Omniscio) Omniscio's Slack integration pulls selected channels and DMs into the unified inbox so you can read, triage, and reply to Slack messages without switching apps. It connects to your workspace through Slack's Socket Mode (a WebSocket connection — no public webhook URL needed), so messages arrive in real time. Telegram integration (read and reply to Telegram chats from Omniscio) Omniscio's Telegram integration signs in with your real Telegram account rather than a bot, so every chat, group and channel you already see in Telegram streams into the unified inbox in real time and you reply inline as yourself. Covers where it lives in Settings, connecting, triaging, the agent-facing CLI routes, and the approval gate on anything outbound. Telegram via mtcute engine (the default Telegram MTProto engine) Omniscio's Telegram integration talks to Telegram's MTProto user-account API, so you sign in with your real Telegram account, and the client library driving that connection is called the engine. Since the F002 go-live the mtcute engine is the default, with the legacy GramJS engine kept available as a reversible opt-out. AI Providers Anthropic outage alerts (API status monitor) Omniscio watches Anthropic's public status page in the background and tells you when an Anthropic incident is what is breaking your Claude sessions — the inbox card, silent toast, in-chat note and all-clear, how it detects incidents, and how to turn the alerts off. Anti-Gravity Provider Anti-Gravity (agy) is a selectable provider alongside claude, codex and gemini — how to enable it, the dedicated virtual project and per-project default, the model story, the gap and error states, and how each turn runs under the hood. Cheap / utility LLM routing The cheap, hardcoded model Omniscio uses for its small utility AI calls such as reply chips, briefings, TTS summaries and voice intent, the four model pickers that remain configurable, the silent Groq to Anthropic fallback and circuit breaker, and the steps to add a new model to the catalog. Codex provider (OpenAI CLI-backed sessions) Codex is OpenAI's own CLI coding agent, and Omniscio can run sessions on it instead of Claude. This page covers what a Codex session looks like, how to switch Codex on, how to sign in with a ChatGPT plan or an API key, and how Codex sessions behave alongside Claude ones. Codex provider (OpenAI CLI-backed sessions) (part 2) Part 2 of the Codex provider page: how Omniscio holds and manages your Codex sign-ins, how a project remembers which engine its sessions start on, and how you pick the OpenAI model and the reasoning effort a Codex session runs with. Codex provider (OpenAI CLI-backed sessions) (part 3) Part 3 of the Codex provider page: what actually runs when you start a Codex session, how Omniscio talks to it and streams a turn back, and where the pieces live in the codebase. Cursor Provider Running Omniscio sessions on Cursor's own CLI agent instead of Claude — a different agent and a different model, not Claude in another harness. Covers what changes on screen, the three readiness gates you have to clear before it can be used at all, choosing from the curated model list, the errors you may hit and what each means, and how a turn is actually run. Cursor Provider (part 2) Part 2 of the Cursor Provider page: how a turn actually runs under the hood — the cursor-agent process Omniscio launches, how its output is turned into an ordinary Omniscio session stream, and the stream format as validated against a real capture. Part 1 covers what changes on screen, the readiness gates, choosing a model and the error states. Custom Providers (bring your own web AI provider) Custom Providers lets you add almost any web AI provider yourself — a display name, a base URL, an API key and the model ids it serves — and run it as a named provider inside the session model picker, in either OpenAI-compatible or Anthropic-compatible wire format. Covers adding one, the URL validation applied, and the built-in OpenRouter entry. DeepSeek Harness Provider DeepSeek Harness as a session provider: DeepSeek's OWN agent harness CLI (`dsh`) driven over the Agent Client Protocol, how it differs from the existing DeepSeek model provider, how to install the CLI and what gates a spawn, and what it does and does not do in v1. DeepSeek Provider DeepSeek as an alternative session provider: how to switch it on and add a key, which V4 models you can pick and what each is good for, the two readiness gaps and their click-to-fix, and where DeepSeek's spend and context-window limits show up. Devin Provider Cognition's Devin as a selectable session provider: how to enable it and add an API key, what a Devin session looks like in the app, the first-message length limit, how cost is reported, and what v1 leaves out. Gemini Provider Google's Gemini CLI as an alternative session provider: how to enable it and add a key, which models you can pick, the readiness gaps and their click-to-fix, and the image, tool-marker and streaming behavior that differs from Claude. GLM Provider Zhipu's GLM as an alternative session provider: enabling it and managing accounts, the naming split between the account UI and the model picker, the model choices, the readiness gaps, and the option to run on company credits. Grok Provider xAI's Grok Build CLI as an alternative session provider — a different agent and a different model, not Claude in a different harness. How to enable it, which models you get, the readiness gaps, and how it reads your skills library. Hermes Provider The Hermes CLI from Nous Research as an alternative session provider — a different agent on your own key, not Claude in a different harness. How to enable it, the readiness gaps, and how it behaves against the other providers. Kimi Code Provider Moonshot's own Kimi Code CLI (`kimi`) as an alternative session provider — a different agent on its own login, distinct from the Kimi model run through the Claude CLI. How to enable it, the two readiness gaps, what reaches the agent (images, your MCP servers) and how it compares with Hermes and Gemini. Kimi Provider Running Omniscio's sessions on Moonshot's Kimi instead of Claude. A Kimi session looks and behaves like a Claude one — same sidebar, same streaming, same approval flow — because the same underlying CLI does the talking; only the model and the upstream endpoint differ. Covers enabling it, choosing models, and what it costs. Meta provider (run sessions on Meta's Muse Spark API) Run Claude-Code-style sessions backed by Meta's Anthropic-compatible Muse Spark API, as an alternative to your Claude account. A Meta session looks and behaves like any other — same approvals, same modes, same cost tracking — and is set up with an API key rather than a sign-in. MiniMax provider (run sessions on MiniMax's API) Run Claude-Code-style sessions backed by MiniMax's Anthropic-compatible API, as an alternative to your Claude account. A MiniMax session looks and behaves like any other — same approvals, same modes, same cost tracking — and is set up with an API key rather than a sign-in. New model alerts (the app notices a new model for you) AI providers ship new models constantly and the picker is a hand-curated list, so the new-model watcher checks the providers you actually have keys for and raises one inbox card when a genuinely new model appears — with a button that starts a session to add it properly. Free, and it only ever alerts once per model. OpenClaw (run agent sessions on a remote gateway) An alternative agent provider that runs your sessions through a WebSocket connection to a remote OpenClaw gateway server instead of spawning a local Claude CLI. Bring your own gateway — you supply the address and an auth token — and the sessions show up in Omniscio like any other. OpenCode provider (run sessions in OpenCode's harness) Run Claude-Code-style sessions through OpenCode, a separate coding-agent CLI that talks to whatever model you point it at. Set up with the binary plus your own provider keys; there is no dedicated sidebar project for it, so an OpenCode session lives inside a real project like any other. OpenRouter (one key, hundreds of models, no Claude account) The connect-with-one-key run path: paste a single OpenRouter key and run real sessions against hundreds of models, with no Anthropic account and no sign-in. The same key also powers the app's built-in AI helpers, and signed-in users can instead run on Omniscio's company key and draw down their prepaid credit. Pi provider (your chosen model, in Pi's harness) A local coding-agent CLI that talks to model providers directly, so a Pi session is your chosen model in Pi's harness rather than Claude in a wrapper — including a local model running on your own machine through Ollama, free and offline. Set up with the binary plus your own keys; text-only in this first version. Porting an OpenClaw setup into Omniscio Bring an existing OpenClaw gateway into Omniscio with as little typing as possible: paste whatever address or token you already have and the fields fill themselves in, then optionally have a Claude session recreate your scheduled jobs and agent personas here. Nothing ported ever runs on its own, and secrets never travel. Provider outage alerts (multi-provider API status monitor) Automatic outage alerting for the API providers you actually use, not just Claude. A provider you have touched in the last seven days is polled and raises one inbox card plus one desktop notification when its API has an incident; a provider you never use stays silent. Qwen Provider (Alibaba DashScope) How to add Qwen as a session provider in Omniscio — enabling it behind the alternative-AI providers toggle, choosing between the five DashScope models, making it a project's default, and what each readiness gap and error state means. Shared chat engine for non-Claude sessions (unified turn engine) Shared chat engine for non-Claude sessions makes Codex, Pi, Gemini, OpenClaw and the other non-Claude engines store and display a conversation the same way Claude does, so their chats look consistent. It has no switch: since 2026-10-02 the shared engine is the only path. Switching between AI providers Omniscio supports multiple AI providers — you are not locked into a single backend. There are two independent provider concerns that control different things: Session provider — which CLI or agent engine runs your coding sessions (the chat where Claude reads files, writes code, runs commands). Options include Claude Code, Codex, Gemini, DeepSeek, Kimi, Kimi Code, OpenCode, Cursor, Grok, OpenClaw, and more (see the full table below). Plugins & Marketplace 33 The plugin system, the plugins that ship with it, and where to find more.
Plugin System Documents Permission (what a plugin must ask before touching your writing) The permission a plugin must ask for before it can touch the documents you wrote in the app. Two levels — read, and create/edit/delete — what each covers, why the AI Writer now asks you to re-approve it, and what is still not open to third-party plugins. Plugin AI-Native Sessions The in-plugin session experience: start and open AI sessions inside any marketplace plugin, with a Sessions pane, suggested-prompt chips, first-class visibility in the Inbox, an auto-context briefing the agent receives, and optional standing context contributed into your own projects. Plugin Bridge Capabilities (part 2) The second half of the plugin bridge: reading the user’s past session history, brokering files and commands in a granted project, picking single binary documents, structured model output, speech-to-text, sharing, spend totals and inbox cards. Plugin Bridge Capabilities (what a plugin can call) Every capability a plugin can call through Omniscio’s sandboxed bridge: the permission model, and the host-primitive and session methods, with what each one refuses and the bounds it enforces. The data, file and output capabilities continue on part 2. Plugin Capability Card When a marketplace plugin can be driven by AI, a one-line Capability Card appears at the top of that plugin's Sessions pane saying so. It is read-only, dismissible per plugin, and never runs an action itself — anything sensitive still pauses in your inbox for your approval. Plugin CLI Discovery (using plugins from any session) A way for any session running inside Omniscio to discover which marketplace plugins you have installed and drive them through their CLI endpoints, so an agent recognizes a plugin by name instead of telling you the tool does not exist. Covers discovery, calling actions, and approval. Plugin Developer Mode (load unpacked plugins) Developer Mode is the Plugins setting that loads your own unpacked plugin straight from a source folder on disk — the Chrome-style "Load unpacked" model, with hot reload, dev diagnostics, storage reset and a kill switch, and no packaging or marketplace submission. Plugin Settings Panel (a plugin's own settings screen) A plugin can ship its own settings screen instead of only generated controls. Its card gains a Settings button that opens that plugin's screen full-width, with the declared fields still rendered above it, a back arrow to the plugin list, and no new permissions. Plugin UI Contributions An opt-in plugin contribution: a plugin can add its own buttons to the header toolbar and a session's right-click menu, and navigate the app to a session, project, inbox or virtual project. The buttons are read-only chrome until a click lands back in that plugin's own pane. Plugin Widgets A plugin widget is a small live indicator a plugin can show in Omniscio's titlebar — a count, a score, a status dot — with a popover holding the plugin's own page. Covers the five shapes, the colours a plugin may pick, what a plugin cannot draw, and how to turn one off. Shared Sign-In (one Google login across the plugins you allow) A plugin can ask for Shared Sign-In, which puts it on one shared browser session with every other plugin granted the same permission, so Google's account picker is pre-filled instead of blank. Covers the trade it makes, the limits stated on the card, and how to sign out of them all. Bundled Plugins Beautify Site A bundled skill that takes a bland or obviously machine-made page and turns it into something deliberate, in four ordered steps: pick a direction, fit components to it, add motion, then audit the built page and strip every AI tell. Covers what it refuses to ship, how bold it goes by default, what it costs, and where it lives. Books (built-in plugin) The built-in Books reader for EPUB and PDF books you own: importing a book onto the shelf, the paginated EPUB reader and its reading settings, the separate PDF path, where a book's data and reading position are stored, and how the plugin is changed. Flashcards Spaced-repetition flashcards for learning and retention, shipped as a plugin with its own sidebar tile — create packs, review cards on an FSRS schedule, import from Anki, and watch streaks and stats build up, all stored locally. JLS SOP Assistant The JLS SOP Assistant is a built-in plugin for JLS Trading Co members only. An employee asks a question, it finds the matching SOPs in a shared company library, reads them from Google Docs and Sheets with the employee's own Google access, and an agent on that employee's own Claude account answers, formatted, showing the SOPs it used. Conversations are saved on the employee's computer, take follow-up questions, and each lookup leaves a short summary behind. LMS (Courses) Omniscio's learning-management surface, shipped as an in-repo plugin: a dashboard for whoever runs a course, editors for courses and lessons including an AI outline generator, a learner home and lesson viewer with quizzes, progress tracking with a certificate, and an importer for an existing Docebo catalogue. Ollert (a full kanban tool built into the app) A complete kanban project tool — workspaces, boards, lists, cards, drag-and-drop, labels, members, checklists, comments, due dates and alternate views — brought into Omniscio as a native sidebar feature rather than an embedded web page. Off by default, and signed in with your existing Omniscio account. Quick Music Recommendations (new music from the songs you already like) A bundled recipe that turns your liked-songs export into genuinely new music, delivered as no-login YouTube playlists per genre on a page you can open on your phone. A red-team quality pass runs before delivery, refusing anything you already own — running on Node alone, with no Python and no paid API for the default path. Whiteboard (bundled plugin) Whiteboard is a bundled, off-by-default plugin that gives you a library of infinite sketch-and-diagram canvases built on self-hosted Excalidraw, with an AI agent able to read a board and draw on it through Omniscio's local control server while your open canvas updates live. Covers opening it, drawing, linking a board to a task, the board permission rules, and what this version does not do yet. Marketplace Calendar Scheduler (Plugin) Calendar Scheduler is the free Omniscio marketplace plugin that turns the app into a Calendly-style booking tool: define event types and availability, publish a public booking link, and let people book straight into your Google Calendar while Omniscio handles approval, reschedule and cancellation — and, by default, puts today's meetings in your inbox. Forms (Plugin) Forms is the free Omniscio marketplace plugin for asking people questions: build a form, publish it as a public link anyone can open without an Omniscio account, and collect the answers back in the app — where an optional auto-send rule can hand a new response to an agent session. Foundry (AI-powered app builder) Foundry, the Marketplace plugin that carries a one-paragraph idea to a planned, built and optionally deployed app: the thirteen-step PRD interview, Autopilot, the Red Team review passes, the dependency-ordered multi-session build engine, the Firebase deploy path, and what each paid stage costs. Foundry is marketplace-only + how plugins update Foundry, the PRD-authoring plugin, is no longer bundled with Omniscio — you install it from the Marketplace like any other plugin. What changed for existing users, how plugin updates work, and why the migration cannot lose your data. Pattern Oracle (Plugin) Pattern Oracle is the free Omniscio marketplace plugin that reads your pull requests and sessions over a rolling window and writes a periodic intelligence report: the recurring patterns, errors and optimisation opportunities it found, ranked, with recommendations you can track. Plugin Marketplace Omniscio’s built-in plugin catalogue: what the Marketplace is, where to find it, and how browsing, installing, uninstalling and updating work — the auto-update check, the version compatibility rules, the gate on unverified third-party plugin backends, and the approval card an agent’s install request raises. Plugin Marketplace (part 2) The second half of the Plugin Marketplace page: what a plugin may read from your session history and boards, what it can watch you doing, ratings and reviews, the errors you can hit, the developer and publisher surfaces, package signing, and the registry and install internals. Reading Queue (Marketplace plugin — save links, auto-summary, read aloud) The Marketplace-only Reading Queue plugin: saving links, pasted text and PDFs, AI summaries you can highlight, note and search, and read-aloud — plus the acceptance checklist to re-check after any change to its UI. RepoGuard -- Repo Health & Security Scanner RepoGuard is an Omniscio plugin that scans repositories for security vulnerabilities, dependency issues, configuration problems, and code quality concerns. It produces a composite health score (0-100, letter grade A-F) from 7 scanner categories and generates actionable findings with fix suggestions. Share automations (the automation marketplace) The Automations tab of the Marketplace — sharing a recipe as a copy-paste code with the private parts stripped out, publishing it to the curated cloud catalog, importing someone else's automation as a paused recipe, and the publisher gate that guards it. Storyboard Planner (pre-production storyboard plugin) Storyboard Planner is a first-party plugin that turns a video script into a shootable storyboard constrained by your real crew, gear, locations, and talent. It asks about your production setup upfront (or loads a saved Vault profile), breaks the script into shots, flags any shot that exceeds your stated rig, generates per-frame images via Image Studio with character reference photos for consistency, and publishes a shareable storyboard link. Every frame is something you can actually shoot. Typing Tutor Typing Tutor is a marketplace plugin that teaches touch typing from the home row up, adapts its practice text to the keys you are slowest or least accurate on, and wraps it in a light game layer of XP, levels, streaks and achievements. Everything runs locally with no AI, no network and no cost, it is desktop-oriented, and progress from the old built-in version carries over. Virtual Pets (Plugin) A cosmetic desktop companion: a pixel-art pet that roams your Omniscio window and reacts to what your sessions are doing, from celebrating completions to worrying about errors. It changes nothing about your work, all of its state is ephemeral and resets on app restart, and since May 2026 it ships as a free marketplace plugin rather than a built-in feature. Workspace automations (your company's private catalog) The Workspace tab in the Marketplace is your company's own private automation catalog — the recipes that mention your own projects and accounts, shared with teammates instead of the public. Covers publishing one, taking one down, and the policy a workspace admin can set over what your team should have installed. Cloud & Teams 56 Work that leaves this computer — cloud sessions, teams, sharing and other devices.
Cloud Sessions Letting an agent run its checks here (cloud-enforce window) A bounded window, granted from an ordinary approval card, in which one agent may run its heavy checks on this machine while the testing setting (cloud first) would otherwise send them to a remote test machine. It covers one agent by default, always expires on its own, and never turns enforcement off for anyone else. SSH Remote (run sessions on another machine) SSH Remote lets any individual Omniscio session run its Claude CLI on a different machine — your GPU box, a Linux dev server, a cloud VM — while you drive it from the Omniscio UI on your desktop. Under the hood Omniscio opens an SSH connection to the remote host, pipes a bash bootstrap script over stdin, launches the CLI remotely, and streams the NDJSON output back to your local UI as if the process were running here. Starting a cloud session from a saved copy (project templates) A cloud session normally copies your whole project onto a fresh machine and installs its dependencies again before the agent can start. Once a project has run one session, the app saves that starting point under a name and later sessions begin from it, sending only what changed. The first session of a project is unchanged, the saving happens on a machine of its own in the background, and you can switch the whole thing off. When a cloud session won't start A cloud session uploads your project to a fresh machine before the agent can start. The app now works out how big that upload is before it rents anything, sends one project upload at a time, says what the upload is doing while it waits, and tells you plainly why it stopped when it does. If a session cannot start, the message names the real reason and the way out instead of leaving you to guess. Your cloud session's work comes home and lands itself Work a cloud session brings home is tested and landed by the app itself, so you never run ready-to-merge on it by hand. Four things never happen automatically: landing a dependency change, landing what a secret check could not clear, marking a branch ready off a check that missed its files, or re-trying an answer that cannot change. A project's ignored `.env` files can be delivered separately, and stopped disks or whole-tree captures can retain secret files. Team Chat Agent lanes — letting your agents talk to a teammate's How your agents and a teammate's agents message each other hands-free over a private lane in Team Chat: one step to reach a teammate's agent, answers that come back to the agent that asked, a list of threads under each teammate, approving teammates one at a time or all at once, and limits only if you set them. Channel Invitations (per-channel roles + who-can-post, in Team Chat) Per-channel roles and posting permissions inside Team Chat — the Slack-style model where you choose who is in a channel, whether each person is a Manager, Member or Read-only, and whether everyone or only managers may post. Covers where the Members panel lives, what each role can do, and how to add, promote or remove the people in a channel. Cross-Org Connections (Team Chat Lane C) Cross-Org Connections is the Team Chat lane that starts a one-to-one DM with any Omniscio user by exact sign-in email — a coworker, or someone in a completely different organization — with the connection forming only once the recipient approves. Covers where to send and approve a request, and the privacy switch that opts you out. Slack Import Wizard Status: In-development (gated behind slack-import unreleased feature flag). A guided, multi-step wizard that imports a Slack workspace's channel history into Team Chat. The wizard discovers a connected Slack workspace, lets the user pick channels, reviews the Slack-to-Firebase user auto-match, then runs the import with live progress and a final report. Slack Workspace Import (migrate your Slack into Team Chat) Status: In-development (gated behind slack-import unreleased feature). Reveal via Settings -> Lab -> Slack Workspace Import, or launch with AMCSHOWSLACKIMPORT=1. Team Chat (built-in channels and direct messages) Team Chat is Omniscio's built-in team messenger — channels, direct messages, threads, reactions, pins, search and file sharing, delivered live to the desktop app and your phone, plus self-serve workspaces you create yourself and share by invite code. This page covers what it does, where to turn it on, how it behaves, and the engineering behind it, across a parent page and three more parts. Team Chat (built-in channels and direct messages) (part 2) Part 2 of the Team Chat page: the second wave of features — day dividers, drafts, saved items, custom status, scheduled send, attachments and custom emoji — plus the workspace story, with per-channel roles and invitations, self-serve workspaces you create and share by invite code, and 1:1 connections with people in other organizations. Team Chat (built-in channels and direct messages) (part 3) Part 3 of the Team Chat page: the phone client and the surfaces that sit around the core chat — the standalone phone web app and its push notifications, notification preferences, rich link previews, the project-management bridge, share-comment cross-posting, starting a Claude session from a thread, the Session Starter Bot, the media gallery and video-call backgrounds. Team Chat (built-in channels and direct messages) (part 4) Part 4 of the Team Chat page: what the feature deliberately does not do and the shortcuts still to come, then the engineering — how messages, threads, pins and reactions are stored and kept honest, how search is built and why it is not switched on yet, and what an operator still has to do to take Team Chat live. Team Chat API Platform (planned) Status: In-development (architecture documented, no OAuth2 layer built yet). Registered as unreleased feature team-chat-api-platform. Team Chat dead-letter recovery (re-file failed scheduled messages) Status: Server-side Cloud Function is deployed and complete. No UI surface exists yet (in-development, no feature flag). Team Chat history export (planned) Status: In-development (not yet built). Registered as unreleased feature team-chat-export. Allow users to export their Team Chat channel or DM history to a portable format (JSON, CSV, or both). This is a GDPR Article 20 data portability requirement for a multi-user cloud product. Team Chat hover toolbar extensibility (future plugin/integration hooks) Status: Planned (no code yet). Depends on a broader plugin/extension framework. The message hover toolbar (TeamChatHoverToolbar.tsx) has a fixed set of actions: Add reaction (emoji picker) Reply in thread Edit (own messages only) Overflow menu (pin, save, copy, quote, delete) These are hardcoded in the component. There is no plugin hook or dynamic action registration. Team Chat internal-link warning (a link only you can open) When you type or paste a link into a Team Chat channel or DM that only works on your own computer, Omniscio shows a small amber notice under the message box warning that teammates cannot use it. It never stops you sending, it disappears on its own once the link leaves your draft, and for some links it offers a one-click way to send something your teammates can open. Team Chat message forwarding (forward / share a message to another channel or DM) A planned Team Chat feature that forwards, or shares, any message from one channel or DM to another, the way Slack, Teams and Discord let you reshare a message across conversations. The forwarded message would arrive in the target channel attributed to the forwarder with a visible link back to the original, and the feature is gated and not yet visible to users. Team Chat message translation (inline AI-powered translation) A planned Team Chat feature that translates any message inline, on demand, using the app's existing AI call layer. A Translate action on the message hover toolbar or the mobile long-press sheet would send the message text to the model and show the translation below the original, into the language you have already configured. It is gated and not yet visible to users. Team Chat summarization (planned) An in-development Team Chat feature that will add AI-powered summarization to channels and threads, through a Summarize or Catch me up button in channel headers and thread panes, plus channel summaries, thread summaries and a one-click catch-me-up that respect channel membership and visibility rules. Team Time — World Clock Dashboard Team Time is a built-in world clock for a team spread across timezones: live clocks with working-hours status dots, a meeting planner that shows a proposed slot in everyone's local time with overlap detection, and a reverse converter for scheduling emails into someone else's workday. It is off by default, behind a Settings toggle, a toolbar Globe icon and a Quick Launch tab. User Management An owner and admin console for running an Omniscio organization: inviting people, reviewing requests to join, setting roles and subscription tiers, disabling or deleting accounts, and managing Team Chat workspace membership. It belongs to the optional cloud identity layer, so it is absent on a single-user install, and the desktop console ships only in internal builds. Video/Voice Calls in Team Chat (LiveKit WebRTC, in development) Video and voice calls inside a Team Chat channel, powered by LiveKit: start a call from the channel header, ring the other members, and talk peer-to-peer with up to fifty participants, with mute, camera, screen share and virtual backgrounds. Still in development behind a Lab flag, with the cloud functions and the data processing agreement outstanding before it ships. Workspace icon color Every self-serve Team Chat workspace tile in the switcher rail can be given a custom gradient colour so you can tell your workspaces apart at a glance, from a palette of eight. The colours are stored locally on each device rather than synced, and the company workspace keeps the standard building glyph. Mobile & Devices Cross-device agent messages An agent on one of your computers can message an agent on another, when both are signed into the same Omniscio account. Your computers find each other on their own, and you connect two of them with one click on EACH machine. Each conversation is a named thread pointed at a session on each computer, so several run at once; every conversation starts muted and out of Team Chat. Cross-Device Sync Sign in on a second computer and your setup is already there — sessions, projects, recipes, automations, bookmarks and non-secret settings, kept end-to-end encrypted. Covers the opt-in switch, device enrollment and the one-time recovery code, what happens to running sessions and scheduled jobs, and the security guarantees. Instant new session on mobile (create-on-send) On a phone, tapping New Session opens a real new-session screen instantly — the same setup controls you get on desktop, with a composer and no spinner — and the session is only actually created when you send your first message. On by default; covers what you see and why it cannot leave a ghost session behind. Mobile Access (run and triage your sessions from a phone) Reach your Omniscio install from a phone: what you see when you open it, how you get in, and the first things to know before you rely on it. The two parts below cover staying connected when the network misbehaves, and what to do when something does not work. Mobile Access — staying connected (part 2) Part 2 of the Mobile Access page: what the phone shows when the connection is poor and how it recovers — the offline and reconnecting indicators, how fast a page resumes, the recovery that runs when the app looks stuck, the transport that survives a freeze, signing in on a phone, and the settings behind it all. Mobile Access — when something is wrong (part 3) Part 3 of the Mobile Access page: the troubleshooting walkthrough for a phone connection that will not behave, followed by the internals — how the mobile transport, session handling and resume actually fit together for anyone working on the code. Mobile back gesture (Back steps back, it does not exit the app) On a phone, the system Back gesture steps back one screen inside Omniscio instead of closing the app — Android's Back button or swipe, and the iOS edge-swipe in a standalone-installed app. Which screens catch it, how far back it goes, and the two platforms' differences. Mobile device tracking (which phone is connecting, and should it) Your phone reaches Omniscio with a pairing token, and anything holding that token used to be able to connect with no idea which device it was. Now a new device is named and surfaced, you can block one without re-pairing everything, and a device has to be approved before it can do anything at all. Mobile network priority An active authenticated phone receives the highest connection priority across Omniscio-managed network activity. It is always on and has no setting to configure, and it applies to both the normal off-loop mobile transport and its in-process fallback. Mobile performance (making the phone app open fast, and stay fast) How the phone app is kept quick to open and quick to use: an app-shell cache so a reopen is instant, a warmed payload instead of a cold fetch, and budgets that exist to be watched rather than guessed at. Also covers the optional checks that catch a broken or non-booting mobile build before you find it yourself. Multi-select on mobile (long-press a row, act on many) On a phone, press and hold a session row to start selecting, tap more rows to add or remove them, then run one bulk action from the bar that replaces the tab bar. Covers the 500 ms long-press, how the selection is dismissed, and why the gesture occasionally does not fire on iOS Safari. Phone Control (drive an Android phone from Omniscio) An in-development feature that lets Omniscio drive a connected Android phone over the local Android Debug Bridge — a Phone panel with a live screen, plus agent tools any session can call. Android only, desktop only, off by default, and revealed from Settings → Lab. Swiping a mobile pop-up away (flick down to close) On a phone, flick a pop-up down to close it — the account panel or any of the app's bottom sheets — using the small grab handle at the top when you want it to close straight away. Which pop-ups it applies to, which it deliberately skips, and why the desktop app is unaffected. Tap the project title to jump to the top (mobile) On a phone, tapping the project's name at the top of the screen scrolls its session list back to the top — the familiar tap-the-title-bar gesture. It applies to a real project's session list on the narrow touch layout only, and tapping the same title while you are reading a session still steps back out to the list. Your phone reopens the session you had open (mobile) Refresh the phone page or reopen the app while you have a session open and Omniscio puts you straight back into that same session, painted instantly, instead of landing you on the projects list. Covers when it applies, when it deliberately does not, and what you see. Sharing Artifact Sharing (shareable HTML of sessions, messages, selections) Turning any session, message or highlighted selection into a self-contained public HTML page: how to share one, where the Shares list lives, what the generated page contains, and how expiration, passwords and view limits are enforced. Branded report template (agents deliver ready-made report pages) When an Omniscio agent produces a report deliverable — research findings, a tool roundup, recommendations, an audit, a comparison — the product renders it server-side from plain data into one standard branded report page, so every report published to Shares arrives on-brand and consistent. Custom share URL name (custom slug + short link) Paid accounts can give a share link a memorable name instead of a random token, so it reads as a word rather than a 64-character string, and can opt into a short link on top. Covers who is eligible, the naming rules, what happens when a name is taken or re-used, and how the in-app viewer resolves either kind of link. Reporting a shared page, and taking one down Reporting a published Omniscio share, and taking one off the public web: the "Report this page" link every public share carries, what a report contains, who sees it, and the owner-only switch that removes a page without deleting it. No page is ever scanned automatically. Share a session with a teammate (the owner half of Multiplayer Sessions) Share this session invites a named teammate into one of your live sessions. They read the conversation in their own Omniscio under Shared with me, and — if you grant it — reply into the same conversation under their own name. This is the sending end. The receiving end is shared-with-me.md. Both halves are one feature (multiplayer-sessions), off by default, and entirely separate from the unrelated Shares feature that publishes artifacts to a public URL. Share Artifacts (publish local files and pasted content as links) Share Artifacts turns a local file on your machine, or arbitrary HTML / Markdown / Text / React you paste into a modal, into a public shareable URL in one click. The destination is a self-contained HTML page hosted at https://shares.omniscio.com/s/<token> — viewers don't need Omniscio, an account, or any extension; Share Artifacts (publish local files and pasted content as links) (part 2) The opposite of "Republish as new": keep the same public link but swap in new content. Use it when you regenerated a report at the same path, or want an already-shared link to reflect an edit without re-sending it — viewers of the old URL automatically see the new version. It reuses the share's existing token via the service's republishToken → updatedInPlace path (the same in-place mechanism KMS/Vault page-publishing uses, generalized to any artifact share). Share Artifacts (publish local files and pasted content as links) (part 3) The publish pipeline is in src/main/services/share-artifact/share-artifact-service.ts — single chokepoint for both the renderer IPC (SHAREPUBLISHARTIFACT) and the CLI POST /share/publish. Both routes pass through createShareArtifactService({ publisher }) so they share one storage publisher. Share Comments (inline and general commenting on shared links) Viewers of an Omniscio Share can leave comments directly on the share page. Five comment types exist: General comments — free-form, not anchored to any content position. Inline text-selection comments — the viewer selects text on a markdown / plain-text / code / SVG share, and the comment anchors to that selection. Coordinate-pin comments — on image and PDF shares, the viewer clicks a point on the content and the comment anchors to those x/y coordinates (rendered as a numbered pin). Share view events (privacy model + audit log for the view-tracking pipeline) When you protect a share (password / view cap / notify-on-view) or even just publish one and leave it open, Omniscio needs to know whether — and how often — strangers on the internet are actually opening your link. That bookkeeping is done by the view-event pipeline: a Firestore-backed audit trail that records one row per view, mirrors a per-share view counter back to the desktop app every 60 seconds, and fires an OS notification the first time each new view lands if the share has notify-on-v... Shareable page links (Copy link to this page) Omniscio lets you copy a https:// link to a specific app page — the Inbox, a top-level panel, a settings pane — and send it to a teammate. When they open it, Omniscio navigates them straight to that page on their own machine. Shared with me (the invitee half of Multiplayer Sessions) Shared with me is a sidebar entry under Communication that lists sessions a teammate shared in — never your own. Opening one shows that conversation read-only, live, in your own Omniscio; if the owner granted you can-reply, a composer appears at the bottom and your messages go into their session attributed to you by name. It is the receiving end of the owner's Share this session dialog. Shares view (master/detail sidebar + in-app webview viewer for shared links) Shares is a top-level entry in the Omniscio sidebar (under the Omniscio group, alphabetically between Settings and Skills) that opens a Cron-Jobs-shaped master/detail layout: a 268 px sidebar on the left lists every shared link you've created, and the right pane shows the currently-selected share with a three-mode in-app webview viewer (edit / view / fullscreen) and inline controls for label / expiration / view-cap / notify / password / send / revoke / delete / republish. Shares view (master/detail sidebar + in-app webview viewer for shared links) (part 2) Sidebar entry. Registered in INTEGRATIONREGISTRY (left of the existing Settings / Skills rows in the Omniscio group) as a virtual project with sentinel SHARESPROJECTID (from src/shared/virtual-project-ids.ts). UI registry shape. The Shares entry in src/renderer/src/integrations/ui-registry.ts is a two-component integration: panelComponent: lazy(() => import('../features/shares/SharesDetailPane')) + sidebarComponent: lazy(() => import('../features/shares/SharesSidebar')). Team sharing (share skills, automations & saved prompts with teammates) Team sharing hands a skill, an automation or a saved prompt to other people, either as a one-off share code you send however you like, or through a members-only Team Library shelf that everyone in the same workspace can install from in one click. Imports are inert, secret-scanned and size-guarded, and the Team Library is still off by default. Settings & Customization 58 Making the app yours — how it looks, how it behaves, and what each setting does.
Appearance AI Tools group (sidebar) The AI Tools group is a collapsible row in the Omniscio projects sidebar that nests the features you interact with as an AI — AI Council, AI Coaching, Journal, PM Advisor, What Changed and Ops Engine — collapsed by default, with members you can expand, roll up, or drag out for solo use. Animated logo The Omniscio logo swirls with the brand animation in the title bar, Quick Launch, the startup loading card, the closing overlay and the Windows downloader. It plays only while someone is using the window, rests when nobody is, follows reduce-motion, Motion & Polish and Low Power Mode, and turns off in one step: right-click the logo, or flip Animated logo in Appearance settings. Automation group (sidebar) The Automation group in the projects sidebar — a collapsible row that nests Automation Helper, Cron Jobs, Recipes, Automations, Quick Replies and Scheduled Messages in one tidy place, its roll-up panel, and how to move a member out for solo sidebar use. Chat Depth (Flat / Soft / Glass dial for message styling) The Chat Depth setting — a three-way Flat / Soft / Glass dial that controls how much shadow, translucency and top-edge sheen the chat message bubbles, the activity pill and the composer carry. Where the card sits in Settings, what each rung looks like, why the default is Soft, and how it stays independent of whichever visual theme is active. Choosing Which Items Appear in the Sidebar The projects sidebar can show a range of optional items — AI tools (Ask Omniscio, AI Council, Journal, PM Advisor), productivity apps (Scratchpads, The Vault, Mission Control, Ollert, Arij), creative tools (Image Studio), communication (Supermail, Drip, Screen Recording), and more. You pick which of these appear via the Sidebar Items panel. Collapse the sidebars (maximize chat area) Two independent toggles let you hide either side panel on the Omniscio dashboard so the chat takes the whole window. The projects sidebar is the narrow leftmost column (project rows, color bars). The sessions sidebar sits to its right (session rows for the active project, inbox at the top). Either can be collapsed without affecting the other. Communication group (sidebar) The Communication group in the Omniscio projects sidebar — the collapsible row that nests the messaging and meeting tools (Team Chat, Email Cleanup, Email Summarizer, Drip, Meetings, Screenshots, Screen Recording, Voice History, Voiceprint Studio), and how to expand it, open a member from it, and move a member in or out. Creative group (sidebar) The Creative group in the Omniscio projects sidebar — the collapsible row that nests the tools for making and capturing content (Writer, Flowcharts, Image Studio, Decks), and how to expand it, open a member from it, and move a member in or out. Custom Session Groups (create your own collapsible sidebar sections) How to make your own collapsible sections in the left session sidebar — like the built-in NEEDS YOU or PAUSED sections — by flagging a library tag to show as a group: where the controls live, how scope and empty groups behave, and why a group is a label, not a folder. Customize the header toolbar The row of small icon buttons on the right of the title bar, and how to rearrange it — pin a tool into the bar, send it to the "More" overflow menu, or hide it entirely. Covers the right-click menu, the grouped overflow list, drag-to-reorder, dividers, hover descriptions and the full Widgets manager in Settings. Customize the session menu (reorder + hide items) Lets each person reorder the entries in a session's overflow menu and hide the ones they never use, so the menu shows what they reach for, in their order. The grouping into labelled, collapsible sections with a filter box ships for everyone; the customising half is behind an in-development flag that is off by default. Developer Tools Group The collapsible sidebar row that gathers Omniscio's build-and-run developer tooling into one place so the sidebar stays tidy — which tools can appear in it, how to expand and collapse it, and how each member is switched on. Header widgets The named controls in Omniscio's titlebar: clicking one either acts directly or opens a small popover. Which widgets ship, how to rearrange them by dragging, and how a plugin contributes one of its own. Hub Descriptions Omniscio is made of many hubs — built-in tools, sidebar categories and your own projects — and this feature gives each one a short plain-English line saying what it is, with the whole set editable by you. Motion & Polish (the app's visual flourishes) A master switch plus one toggle per visual flourish — the small animations and transitions that make the app feel finished. Everything is on by default, one switch turns it all off at once, an explicit opt-out is never overridden, and every effect quiets itself under your system's reduce-motion setting and Low-Power mode. Per-window zoom (each popped-out window remembers its own size) Every secondary window you pop out — a session, a project, the vault, the scratchpad — gets its own zoom level that scales the whole window and is remembered next time you open it. Session windows remember per session, so two sessions can come back at two different sizes. Pin session commands to the session bar (Snooze, Pause, etc.) Session-bar pinning promotes the actions you use most out of a session's overflow menu and onto its header toolbar as one-tap icon buttons: Snooze, Pause and Unpause, Send Later, Copy app link, Aside Mode and Archive. A pinned command leaves the menu, and unpinning sends it straight back. Archive is pinned by default. Pop out a project or integration window Popping a project, an integration panel, or a marketplace plugin panel out of the main Omniscio window into its own desktop window: what qualifies, where the entry point lives, how the window follows your theme and remembers its position, and how a popped-out plugin differs from a popped-out project. Presentation Mode (recording-safe "show only what I pick") Presentation Mode is the recording-safe screen: one switch that hides your real projects, sessions, account identity, live cost figures and the leaky bits of header chrome, and reveals only what you explicitly pick. Covers the staging panel, the three hide styles, the panic control, and every surface it neutralizes. Productivity group (sidebar) The Productivity group: a collapsible row in the Omniscio projects sidebar that nests the general-purpose tools which fit no more specific group. Covers the thirteen tools it holds, how to expand it, use it, and take a member back out. Projects Sidebar — Default "Omniscio" Group The collapsible Omniscio group at the top of the projects sidebar, which collects Omniscio's built-in virtual projects so they don't clutter your real ones. Covers what is in it, the sub-groups and the Omniscio-versus-Plugins split, the colored status counts, and the migrations. Prompt Tools group The sidebar group that gathers Omniscio's prompt-crafting features — Super Prompts, Super Prompt Creator, Bake-Off, AI Council, Clean Room and Missions — into one place, plus the sessions overflow-menu submenu and per-tool modals that reach them without leaving the current view. Session Auto-Color Rules (keyword-based or project-inherited automatic session coloring) Auto-Color Rules give sessions a colour bar automatically, so you never colour-code by hand. One mode makes every session inherit its project's colour; the other matches keywords against the session title in an ordered list, first match wins. A colour you set manually always outranks both. Session list — per-section sort, filter and Overseer crews Sort and filter each sidebar section on its own, and expand an Overseer to see its live crew. Covers the three orders each section header offers, the filter box and the banner that says how many rows the filter is hiding, and the crew list that folds out under an Overseer row. Show only active projects (funnel filter) A funnel toggle in the projects-sidebar header collapses your project list down to just the projects that have something happening — a running, waiting-on-you, or errored session — and hides the quiet ones. Off by default; flip it on when a long project list is mostly idle and you only want to see what needs you. Sort Projects by Usage An opt-in projects-sidebar display mode that ranks your projects by how much you've used them lately — the number of sessions you have started in the last 30 days — so the ones you are actively working on rise to the top of the group they are already in. It is a display preference — it never renames, moves, archives, hides, or deletes anything, and turning it off restores your manual drag order exactly. System group (sidebar) The System group is the collapsible sidebar row that gathers Omniscio's own app-level surfaces — settings, search, tags, alerts, marketplaces, briefings, stats and shares — into one place. This page covers what sits inside it, how to expand, use and reorder it, and what an agent needs to know about its stable internal id. Text Size (separate Desktop and Mobile text size) Text Size controls how large all text, and the whole UI that scales with it, appears. There are two independent settings, one for the desktop app and one for phones and narrow touch screens, each offering the same seven steps from Extra Small to XXX-Large. Both are reachable from Settings, from Ctrl+plus and Ctrl+minus, from Ctrl+scroll, and from the phone header menu. UI Auto-Tidy Auto-Tidy uses Omniscio's local signal for which controls you have not used recently to quietly declutter the interface over time. After 30 or more days without touching a sidebar integration or a toolbar icon, it moves the item somewhere quieter. Tidying is a reversible demotion, never a deletion, it is opt-in and off by default, and all of it is restorable. Visual Themes (13 bundled skins + custom themes) Omniscio ships thirteen bundled visual skins and lets you import your own theme files, all driven by CSS variables so picking a skin recolours every component without touching code. Covers where the theme grid lives, how to import, switch, delete or write a theme, what a theme can restyle, and the light-mode and accent-contrast rules a theme has to keep. Wayfinding — the breadcrumb trail and back/forward arrows The desktop titlebar carries a location trail plus browser-style back and forward arrows, so you can always see which group the current screen belongs to and how to get back to where you were. Covers where the control lives, how a location is worked out, how to remove it, what the arrows remember, and which panels deliberately publish or skip a third crumb. Behaviour & Shortcuts App Update Notifications (auto-update banner + error self-recovery) How Omniscio updates itself: the silent background download, the Ready, Error and Duplicate-copy banners, manual update checks, rolling back to a previous version, selective rollout, and the What's New card that lands in your inbox after an update. Custom Render Rules (turn your own regex matches into inline badges) Custom Render Rules turn your own regular expressions into inline badges inside Omniscio conversation text — on an agent's output and your own messages alike — with a semantic colour tone and optional icon, label and link built from capture groups. Covers adding a rule, the badge options, the safety limits, and where rules are stored. Focus mode Two independent features share this name. The headline one batches Omniscio's noisy inbox alerts into a single interruption when a focus window ends; the older one is a view-only sidebar collapse. What each does, and what they deliberately leave alone. Focus mode (part 2) How Focus mode is built: the CLI entry points, the rule engine and its state machine, the banner count the main process owns, and the contracts the suppression and overlay layers answer to — the implementation detail behind the behaviour on part 1. Hotkey Training Mode (block the mouse to learn the keyboard) An opt-in mode that helps you build keyboard muscle memory by gently refusing the mouse: the first click on a control that also has a shortcut is blocked, and a bubble tells you the key to press instead. Hub jumper (Ctrl+G — type two letters, jump to any hub) A window that opens in the middle of the screen on Ctrl+G and does two jobs: it shows every hub as a grid of icon tiles to click, and it jumps you straight there the moment you type enough letters to identify one. Keep computer awake while sessions are running (cross-platform) A setting that asks the operating system not to sleep for exactly as long as at least one session is actively running, then lets it sleep normally the moment the last one finishes. It exists because a sleeping computer suspends the running agent mid-turn, silently stalling work that was meant to finish unattended. Keyboard shortcuts (every binding, and how to change them) Every keyboard shortcut Omniscio ships, grouped by where it applies, plus how the binding system works: rebinding any of them, the optional one-handed layout and the hotkey training mode that helps you learn them. The full list is long, so it continues on the parts linked below. Keyboard shortcuts (every binding, and how to change them) (part 2) Part 2 of the keyboard shortcuts reference: the bindings that drive sessions and the message box, and the ones for quick replies and the AI helpers. Continues from the overview page, and the navigation, search and display bindings follow in part 3. Keyboard shortcuts (every binding, and how to change them) (part 3) Part 3 of the keyboard shortcuts reference: moving around the app, searching, changing what is displayed, and the bindings that only apply inside a particular project — Gmail, the diff review view, and the daily digest. Keyboard shortcuts (every binding, and how to change them) (part 4) Part 4 of the keyboard shortcuts reference: how the binding system actually works underneath — what happens when you press a key, how a binding is registered and checked for conflicts, and how the one-handed layout and training mode reuse the same machinery. Language Support (UI & AI replies) Omniscio's interface ships in 35 languages, all available with no setup, and the AI speaks back in whatever language you are writing in. Two independent halves: the UI language you pick in Settings, which switches instantly, and the spoken reply language, which is detected automatically and is always on. macOS permissions — what Omniscio asks for, and why A plain-English reference for the macOS privacy prompts Omniscio can produce: what each permission is actually used for, which ones are on by default, and what to do about the ONE notification that repeats when macOS blocks Omniscio from modifying apps on your Mac. Multi-level Settings navigation (category drill-down) Multi-level Settings navigation changes how you move through Settings: instead of one long scrolling list of every section, Settings opens on a category picker; you pick a category, then a page inside it, with a breadcrumb back to the top. It shipped as the default for everyone, and it also brings the room grouping that reorganises Settings into clearly named areas. New Settings layout (the reorganized rooms) The Settings screen reorganized into clearer, better-named rooms — top-level groups that gather related settings together — plus a two-level drill-down and tabs inside the largest pages. It changes the grouping, labels and order only: nothing is added, removed or re-named underneath, and section links keep working. Quick Launch Modal Omniscio's global-hotkey floating composer: press Ctrl+Space anywhere, type a prompt, and a new Claude session starts in the project you choose. What the modal looks like, how to open it, how to compose and send from its Session tab, and every setting on its own Settings → Quick Launch page. Quick Launch Modal (part 2) The second half of the Quick Launch page: every tab the strip can hold besides the Session composer — Search, Ask Omniscio, Automation, Alarm, Task, Calendar, KMS Note, Find Email and the rest — plus per-repo tabs and the mobile quick-actions sheet that mirrors the same set on a phone. Quick Launch Modal (part 3) The third part of the Quick Launch page, for readers working in the codebase: how the floating window is managed, how its global and per-tab hotkeys register, how each tab maps to a component, the voice-suppression seam that keeps main-window voice commands dormant, and the design decisions behind the modal. Right-Click Setting Affordances Right-click certain live elements of the app to Hide them or jump straight to their setting: the subagent activity card, the Auto-Waited tag, the voice button, the close button, the context-% ring, the alternate-providers chip and the current-branch badge, plus the waiting surfaces, which offer the open-setting menu only. Right-click on desktop, or long-press for half a second on a touch screen. Session title style — choose how titles are written Omniscio names a new session by asking a small AI model to summarise your first message. Title Style lets you choose how that title is written — five built-in presets (Brief, Descriptive, Journey, Latest Topic, Custom), a global or per-project override, and extra instructions you can append. Two of the presets refresh the title as the conversation moves on. Settings (opening it, and getting where you were going) Settings is where you change how Omniscio behaves — accounts, channels, notifications, sessions, keybindings, appearance, voice and the rest — and it opens inside the app rather than in a separate preferences window. Covers every way in, how to get back out, and how to land on one section directly. Settings Home (landing page, changed-from-default, reset) Settings Home is the landing page you see when you open Settings fresh — a one-screen overview that answers where a setting lives and what you have changed. It holds a large settings search box, tiles for the most-used settings, your recently changed settings, a Changed from default list where each row can be reset, a browse-all map of Settings, and quick links. Settings redesign (tabbed pages) The Settings redesign breaks the app's largest Settings pages into tabbed sub-pages, so a long scroll becomes a short labelled tab strip. It also folds the rarely-touched settings on a tab under an Advanced disclosure and adds live previews for a couple of controls. It changes only how the settings look, never what any setting does. Settings Search telemetry A small record of how people use the search box in the Settings panel: what they search for, which result they pick and how far down the list it was, and whether they then change a setting. It exists so Settings search can be made to rank results better, and a read-only tab in Stats shows what has been collected on this device. Settings — Virtual Project Entry Settings is reachable from two places — the gear icon in the toolbar and a Settings entry at the bottom of the Omniscio group in the sidebar — and both activate the same Settings screen. There is no separate full-screen modal, so there is only ever one copy of Settings, and a section picked from either entry lands in the same place. Toolchain Updates (background checks + native-installer notifications) How Omniscio keeps the command-line tools it depends on up to date: a background version check, a one-click Update that runs that tool's own update command in the background on your computer, and the silent npm path kept for the pure-npm tools. Covers the two settings that control it, which tools never update silently, how a stuck install reports itself, and the inbox cards with a one-click Update now when Claude Code, Codex or another recommended tool falls behind. Tray and window The system tray icon and the main window: the two close behaviours controlled by closeToTray, the tray's right-click menu with its show, wake-word, hard-reload, restart, DevTools and quit entries, the two global hotkeys, the single-instance lock, and how popped-out windows and a saved window size come back on the next launch. Account & Billing 28 Your accounts, your keys, and what you are spending.
Accounts & Sign-In Account Indicator (top-right toolbar popover) Everything the account-indicator popover in the top-right toolbar shows: the account list with its usage bars and switch-priority order, the data-driven provider tabs, the Codex and Grok mirrors, mobile non-Claude accounts, metered API-key charges, the re-login tally, and the staleness footer. Account Pool (multiple Claude accounts + exhaustion alert) How Omniscio holds several Claude logins and picks one per session spawn: the tiered default versus the load-balanced opt-in, pinning a preferred account, the all-accounts-at-their-limit banner, dead-login quarantine and failover, and how rate-limited or burst-throttled accounts drop out of the pool. Add a Claude account Adding an account Omniscio uses to talk to Anthropic: the Log In with Anthropic OAuth flow, cancelling a sign-in that is in progress, the single API key with its replace and re-enter paths, activating an account, and what the Settings page and the OAuth flow do under the hood. API Keys (unified key hub) API Keys is the single Agent Tools sidebar row that holds everything key-related: the My API Keys tab for a personal gateway key, and the Stored Keys tab inventorying every key and secret Omniscio can see — names and metadata only, never the values. Choose how your Claude sessions sign in Choosing how your Claude sessions sign in — a managed account pool, your own API key, or a Claude Code login: where the picker lives, what each method stores, and why Managed accounts only appears when it is genuinely available to you. Consumer Terms Gate What happens when an account must accept updated Consumer Terms and Privacy Policy at claude.ai before the API will serve it: instead of retrying forever, the session stops cleanly, says exactly what is wrong, and that account is kept out of the pool so other sessions are unaffected. Covers the four-step fix and why the detection is deliberately narrow. Default Model by Account Plan (account-tier model defaults) The opt-in "Default Model by Account Plan" setting: which Claude model a brand-new session starts on for a Max, Team, Enterprise, Pro or Free Anthropic account, how your own per-session and project choices override it, why the chosen model then stays put, and what it never touches. Deleting your account / starting from scratch Deleting your Omniscio account and starting over: what is removed in the cloud straight away versus cleared on this computer at the next launch, exporting first, the only-owner block on shared workspaces, signing back in afterwards, and the smaller resets that clear one app's data instead. Distribute Sessions Across Accounts The setting that spreads new sessions across all your logged-in accounts by how much headroom each has, instead of filling one account at a time — what it does when on, how to see it working, and how to stop it. Remove a Claude account Removing an account deletes its credentials from this machine and takes it out of the account pool's rotation — it does not sign you out of Claude.ai in your browser or revoke anything at Anthropic's end. Login accounts and API key accounts are removed the same way, from a trash icon on each row in Settings, behind a confirmation. You cannot remove your only account. Sign accounts back in automatically (automated re-login) Driving the whole Claude sign-in for a stale account in your pool: the Sign them back in panel, what the automated flow clicks, the one inbox card raised when a click does not register, the optional Authorize-window setting, the opt-in 3-hourly automatic sweep that repairs a stale login before you notice it, adding a brand-new account, every failure message, and the CLI routes an agent uses. Switch active account There is no traditional "switch account" button in Omniscio — and you almost never need to switch manually. When you have more than one Claude account configured, Omniscio's account pool picks an account for each new session at spawn time. The session header shows a small indicator with the active account's email and a colored capacity dot, so you can always see who is powering the running session, but there is no per-session account dropdown to change mid-conversation. Usage Cascade (what to run when your Claude subscription is out) The usage cascade is a short ordered ladder of other AI vendors that Omniscio falls back to when every Claude account in your pool has genuinely run out, or when the vendor a session is running on (GLM, DeepSeek, Kimi, MiniMax) hits its own usage limit or runs out of balance, instead of parking sessions in a waiting state for hours. You build it in Settings, it ships empty, and a session that left Claude comes home once capacity returns. Usage Forecast (predicted rate-limit exhaustion) Usage Forecast turns your Claude accounts burn rate into plain-English ETAs: a global rollup banner at the top of the Accounts popover, a per-account projection row under each 5-hour bar, the hour-of-day coverage model, and the sessions-remaining estimate. Usage Forecast (predicted rate-limit exhaustion) (part 2) Part 2 of the Usage Forecast page: how to read each per-account forecast state, the threshold that fires an early warning, the data-freshness line, the four-card Usage Stats page, the optional header widget and the architecture behind it all. Who pays & who serves (model vendors) Who pays for a DeepSeek, GLM, Kimi, MiniMax, Meta, Qwen or OpenRouter session, and which company answers it: one ordered list per model family, walked each time a session connects. Covers the settings card, what each row state means, limiting a DeepSeek row to DeepSeek's peak-price hours, what happens when a row runs out of credit, how cloud sessions use your keys without ever receiving them, and how older settings were carried over. Billing & Usage AI spend alerts (cost cards + the opt-in daily spend digest) The inbox cards Omniscio raises when your AI spend looks off — the always-on cost and runaway warnings, the opt-in daily spend digest summarizing yesterday, where each card's actions lead, and which of them a genuine runaway still breaks through. Auto-continue on balance refund (API-key providers) When your own DeepSeek or Kimi account runs out of credit and nothing else in the family's list can pay, Omniscio parks the session in a red out-of-credit state. This in-development feature watches the account balance, continues every parked session the moment you top it up, and reopens the account for every new session at the same moment. Built-in AI helpers (how they're powered — paid cloud vs free BYO key) The small background AI helpers Omniscio runs for you — session titles, suggestions, reply drafts, the daily digest, email summaries — what powers them on each plan, what the monthly allowance actually is, and what the "helpers are paused" inbox cards mean. Buy Credit (prepaid gateway AI credit top-ups via Stripe checkout) Buy Credit lets a signed-in Omniscio user top up their prepaid gateway AI credit with a credit card. You pick an amount, Omniscio opens Stripe's hosted Checkout page in your browser, you pay, and the credit lands on your balance a moment later. It's the self-serve way to keep gateway credit on hand so your AI work never stalls for lack of funds. Buying is 1:1 face value: pay $25 and you get $25 of usable credit. Cost Control Cost Control is the sidebar surface that puts spending in one place — live AI usage and out-of-pocket cost at the top, and every budget cap below it, all editable inline. Covers where it sits, how its caps relate to the Settings screens that own them, the one-time 2026-09-08 historical-spend repair, and how to hide the row. Daily Spend Report (plugin) Daily Spend Report is the bundled plugin that posts one Inbox card a day summarising your AI coding activity — the value the agents produced and your real out-of-pocket spend, read from the live spend ledger. Covers where the card appears, how it self-heals a missed day, and its single report-time setting. DeepSeek Balance Forecast A warning before your DeepSeek API account runs out of money: Omniscio reads the balance every 15 minutes, measures how fast it is actually draining, and raises one inbox card with a Recharge button when the account is projected to reach zero inside your warning window. On by default, reads only, and it fires on a measured drain — never merely because the balance looks low. Kimi balance monitor (low-balance alerts) A background check on your Moonshot Kimi account balance that warns you before the prepaid credit runs out — with a one-click Recharge button — and flags a balance that jumped up since the last check, so a silent auto-recharge on your card does not go unnoticed. Opt-in and off by default. Plan & Usage (your plan, your allowance, your credit) The one customer-facing money screen: the plan you are on and this month's AI allowance, your prepaid credit balance with a top-up button, a Free versus Pro comparison, and the option to bring your own API key. Card payments happen on the payment provider's own page, and a plan is granted by a verified webhook rather than by your browser returning. Pre-send spend warning Warns you before sending a prompt whose estimated cost exceeds a threshold you set: a per-model dollar cap for API-key billing, or a share of the rolling 5-hour rate-limit window for subscription billing. Covers where the thresholds live, when the dialog fires, and the toggle caveat. Stats — Cost, Tokens, Sessions, and Trends Stats is a built-in Omniscio virtual project that surfaces aggregate metrics across every Claude Code session you've run: total spend, token usage, LLM time and active time, per-project breakdowns, and trend lines over time. It lives in the Omniscio group of the project sidebar, between AI Coaching and Settings. Before 2026-04-30 these metrics lived inside Settings → Statistics. Settings is for configuration; Utility AI model (which model runs background AI features) Omniscio runs a small fast AI model in the background for dozens of minor jobs that do not need a full coding session, from session titles and suggestions to daily digests, summaries, and text correction. Out of the box that model is Luna, and adding your own DeepSeek API key silently routes all of that background work through your key instead. Troubleshooting 84 When something is wrong — recovering it, diagnosing it, or making it fast again.
Recovery Aborted-response recovery How Omniscio notices when the Claude CLI abandons its response mid-flight and heals the session itself: the 60-second scan that replays your last message, the two-attempts-per-day cap, the Settings and environment kill switches, the Retry button, load-aware patience, and the v256 upgrade retro-fill. Covers the sibling case too: a reply the CLI reports as FINISHED but that was cut off mid-report is re-asked and held rather than shown (2026-09-29). App is already running, or frozen — restart the stuck copy What happens when you launch Omniscio while another copy is already running: normally the existing window comes to the front, and if that copy has frozen you get a one-click Restart Omniscio dialog instead of nothing at all — plus the safety rules behind it. Auto-restart stuck sessions Auto-restart stuck sessions — the self-healer for a session whose underlying process freezes right after you send a message and spins on "thinking" forever: how the wedge is detected, the setting that turns the automatic restart on or off, and what you see in each case. Backup Mirror (encrypted off-machine snapshot) An encrypted full snapshot of your Omniscio state — the whole database, your config and every attachment — written to a folder you choose, so the install can be reconstructed on a different machine after a hard-drive failure. Backup Mirror (encrypted off-machine snapshot) (part 2) The second half of the Backup Mirror page: first-time activation, the automatic-sync modes, the encryption envelope, the archive's internal layout, retention, the restore safety defences and the IPC channels. Blank-screen recovery button (survives a full React unmount) The Reload Omniscio button that appears in the middle of the screen when the app's interface has vanished, so you can recover without closing the window and losing the forensic trail. Covers when it appears, what it does when you click it, why it survives a React unmount, and how to switch it off. Crash recovery (auto-resume sessions after a crash or restart) What happens when Omniscio starts after a crash or a restart: which sessions qualify, the trickle you watch them reconnect through, how to stop the resume or jump the queue, what to do when a session does not come back, and how to turn auto-resume off. Crash recovery (auto-resume sessions after a crash or restart) (part 2) Part 2 of the Crash recovery page: the crash-loop circuit breaker that quarantines a session, the silent-death watchdog that alerts you the moment the app is killed, the states that are never auto-resumed, why revivals trickle, and where each piece lives in the code. Data folder recovery (your data looks gone after an update) What happens when the app's data folder is renamed between versions, and what to do if you ever open the app and your projects and sessions appear to be gone. Covers where the folder lives on each platform, how the move decides between two real folders, and the Mac sign-in key that has to be migrated with it. Database upgrades & migrations (what happens to your data when you update Omniscio) What happens to your data the first time you run a newer version of the app: the database upgrades itself silently behind the splash screen, a full backup is taken first, and nothing you have written is ever deleted. Also covers rolling back to an older version, the restart-loop guard, and what migrations deliberately do not do. Export / import all data (backup ZIP) The manual, on-demand export and import of your whole dataset as one ZIP — for moving to a new computer, snapshotting before a risky change, or restoring after a reinstall. Covers what the file contains, what it deliberately leaves out, how a full replace is made safe, and the compatibility rules on either side of a version change. Fresh-install starting point (why a new install does not replay the migration ladder) A brand-new install has no database, so it used to build one by replaying every migration — about 1,170 of them, measured at ~9.2 seconds on a fresh machine against 362 ms on an already-migrated one, paid once by every new user. It now starts from a committed, generated image of the finished schema instead, and falls back to the ladder whenever that image no longer matches the migrations the build ships. Git history missing (the store-loss alert) When a project's git history (its .git folder) disappears or is replaced, the app notices within a minute, raises one inbox alert, and stops its own git writes — the auto-lander, worktree cleanup and new worktrees — so nothing is written into a broken or empty history. Starting a session from the alert begins the recovery, and everything resumes on its own once the history is whole again. Is my bug actually fixed? (the release ledger) When you report a bug, a session is started to fix it — and the release ledger answers whether that session actually fixed anything and whether the fix is in the version you are running. It follows the trail from your report to the fixing session to the commits it wrote to the release that carries them, and says so plainly when a report cannot be traced. Model safeguard refusals What happens when a model declines a turn on its own safety system: Omniscio recognises the refusal as its own class, counts it, names the cause in the transcript instead of reporting a generic error, lets the existing automatic retry run (it rebuilds the conversation, which is why most of these heal), and stops a session that is declined twice in a row — parking it visibly with the model and the stated reason. Out-of-context stall recovery When a heavy turn fills the context window mid-generation, an agent can simply stop with no closing message. This page covers the self-healing that follows: Omniscio nudges the agent to continue — or forces a compaction, on engines that cannot compact themselves — up to three times, then hands the session back to you if it still cannot finish. Portable Backup (one-file encrypted export/import) A one-file encrypted backup you carry between computers yourself: what the .amcbackup file holds and how it is protected, how export and import work, how Omniscio picks a full restore or an additive merge, and how the optional credential re-wrap lets your saved secrets travel with it. Post-Restore Credential Wizard What happens when a restore onto a different machine leaves saved credentials undecryptable: the wizard that opens after the restore, the fourteen credential categories it probes, the fix actions it offers per broken item, and the inbox alert and CLI route that surface the same health report. Project folder missing (recreate or relink) What happens when a session's project folder is gone from disk: Omniscio fails the spawn at once and shows a red banner naming the exact path, with a Recreate Folder button, instead of looping on a cryptic spawn error. Covers the three ways to clear it and how automatic recovery waits it out. Relentless session relaunch (keep retrying a launch hiccup until the session is back) When your computer is momentarily too busy to start an agent process, a session that was running can fail to launch for no real reason. Relentless relaunch keeps re-attempting it, with a gap between tries that grows the longer the machine stays jammed, until the session comes back — so a mass restart no longer strands a pile of sessions as red errors. Safe Mode A Windows-Safe-Mode-style boot mode for Omniscio: turning it on restarts the app with all of its automatic, background behavior switched off — automation, scheduled sending, integrations, remote access and self-acting watchdogs — so you can look at what is going on and fix things in full manual control. Session stuck in "needs you" 'Needs you' is not a stuck state — the session has paused itself and is waiting for a decision only you can make. The reason shows as a sub-label on the session dot: Question, Plan Review or Permission. Rate Limited, Auth Error, API Error, Stopped and Recovery Failed are not Needs You — they wait in the project's Interrupted section — and each one has a different fix. Setup Backup to Gmail Weekly encrypted backup of your Omniscio configuration (not your conversations) emailed to your own Gmail inbox, with a one-click restore that pulls a bundle back out of email. Stalled session (orange) and the Continue alert A session turns orange with the label "Stalled?" when its Claude Code process stops producing output for longer than your stall-detection time (Settings → Sessions → Stall detection time, default 15 minutes). It is not an error and not a question — the agent simply went silent. Common triggers: the CLI wedged after a long operation (a big auto-compaction, a slow tool, a dropped network connection). Vendor server fault recovery What happens when a provider's own servers fail mid-request, and why a session that has not yet produced any output recovers by itself instead of asking you to change models. Covers which failures are treated as temporary, which are still treated as permanent, and what each one tells you. Diagnostics Agent Friction A read-only Developer Tools panel over the agent friction ledger — the health line and the ranked list, closing a problem out, the two-hourly triage loop that dispatches fixers under a daily spending ceiling, and how an agent files a report. Console window sentinel (stray black console windows are hidden) The helper that stops stray black console windows from popping over the app on Windows when automated work starts a program with no console to attach to. Covers what it hides, what it deliberately leaves alone, how to read its log, the companion rule about registering scheduled tasks, and the agent-browser equivalent. Export a raw thread (troubleshooting) The raw thread export: a troubleshooting file holding every message a session stored, word for word, each labelled with whether the session window showed it, folded it behind a control, or never drew it. Where to find it, how to read its labels, what it cannot show, and why the file must be treated as sensitive. Heap snapshot diagnostics (memory bloat investigation via CLI) A CLI endpoint that captures a heap snapshot of the renderer, parses it, and returns a small summary of where the memory is going — the supported way to answer "Omniscio is using 2 GB of RAM, where is it going?" without opening DevTools. Hotkey usage A usage signal that records every time a shortcut is pressed and every time you click a control that has one instead — the data behind deciding which shortcuts to keep, which to ship unassigned, and which to surface. Interaction trace (slow-click, jank, and navigation latency log) A single log file capturing every slow click, keypress and rendering stall while Omniscio is open, kept as a continuous tape across launches with a worst-offenders summary at the bottom. It exists so a report of "clicks feel laggy" travels as one attachable file anyone can read. Job Monitor (test/build/dev process dashboard) A read-only dashboard of every long-lived test, build, dev-server and sandbox process Omniscio is holding, plus any runner process the concurrency gates missed — so when a dozen agents work in parallel and one spawns a runaway test loop, you can spot it and kill it. It also shows the wait queue of sessions blocked on a free slot. Job Object orphan-kill (Windows) What happens to the helper processes a session spawns — its tools, its shell — when Omniscio itself dies badly instead of quitting cleanly. A crash, a force-kill or a power loss leaves no chance to clean up, so on Windows Omniscio asks the operating system to tie those processes to its own lifetime and end them with it. Logs & Debugging (where to find logs, export for support) Where Omniscio's diagnostic output actually lives — the authoritative rotating log files and their archive, a separate bootstrap log for the earliest startup moments, and how far back the archive realistically reaches. Covers reading them, exporting a bundle for support, and the automatic crash reporting. Main-process heartbeat (silent-crash forensic trail) A log file recording a continuous five-second tape of Omniscio's main-process health — event-loop lag, memory, live resource counts and the last operation that ran — so that a window which simply disappears with no crash dialog and no error report can still be explained after the fact. Offline banner (is it my network?) A slim amber bar across the top of the app when your computer loses its internet connection, so a failing session or a stalled sync reads as your network rather than Omniscio. Always on with no setting, it clears itself on reconnect — and while it is up your sessions hold quietly instead of burning retries. QW Parser Snapshots (regression-diff baseline) The frozen 18,271-row production snapshot the v1 QuestionWidget parser was diffed against: what it is, why it is deliberately kept stale, how to re-capture it, and how to read the appeared/disappeared diff after a parser change. QW Triage Tool (mine the live DB for parser bugs) The read-only developer tool that scans every agent message in your live Omniscio database, scores each one for likely QuestionWidget parser bugs, and writes a sorted JSONL review queue a human triages — the flags it takes, how the scores are read, and the re-baseline rule. Reclaim disk space (shrink a large Omniscio database) Omniscio's database file only ever grows, and this page covers the one action that physically shrinks it — Settings, Diagnostics, Reclaim disk space — plus what happens at the next launch, the automatic one-time compactions that follow a large cleanup (including the archive of older conversations), the backup and write-log changes that stop the folder and the log ballooning, and why automatic backups are compressed and encrypted. Renderer stall profiler (what the app was doing when it froze) When the app window freezes for a second or more, Omniscio records what its code was doing: a .cpuprofile you open in Chrome DevTools plus one log line naming the heaviest functions. It watches only while the window struggles, never changes how the app behaves, keeps the newest 25 recordings under a size cap, and says why whenever it could not record instead of staying silent. Desktop main window only; turn it off with AMC_DISABLE_RENDERER_STALL_PROFILER=1. Reply Method (how your cursor got into the chat box for each reply) A passive stats tab showing, for every reply you send into a session, how your cursor got into the chat box: you clicked it, you pressed the Reply shortcut, the auto-focus setting placed it for you, or some other way. It lives in the Stats project and is read-only. Only the source is recorded — never your message text. Resources Diagnostic Panel (in-app process monitor) A read-only Task-Manager-style view inside Omniscio showing live CPU and memory for the main process, the renderer, every active Claude CLI session, every running MCP server, every aside agent and any other Omniscio-spawned child process, with a Kill button on the killable ones. Restart attribution (no restart is ever anonymous) In a dev install, agents and automations can restart Omniscio over the command-line API. An unattributed restart looks exactly like a crash, so Omniscio refuses anonymous restart commands and raises a persistent inbox notice naming the session, cron job or supervisor that asked for the restart. Session Event Log A per-session troubleshooting view that lists only that session's lifecycle events: token refreshes, account switches, crash recoveries, pauses, compactions, self-archives and consequential CLI actions, without the chat or the tool output that normally buries them. It answers what actually happened to this session at a glance. Session Trajectory (the full log of what an agent did) Session Trajectory is a read-only full log of everything a session did: every tool call in order, with its full input, full result and how long it took, plus a metrics strip of turns, tool calls, tokens, cost and API time. It opens from a session's menu, or from a sidebar browser that covers recent sessions. It is off by default. Shortcut Efficiency (mouse-vs-keyboard stats with a rough time-savings estimate) Shortcut Efficiency is a passive stats tab that shows, for the handful of common actions that have BOTH a clickable button AND a keyboard shortcut, how often you reached for the mouse versus the keyboard — and a rough estimate of how much time you could reclaim by using the shortcut instead. It lives in the Stats virtual project (the sidebar entry with the bar-chart icon, near Bookmarks / Automations). Startup trace (single-file launch performance log for bug reports) A dedicated log file — startup.log — that captures the timeline of the most recent app launch in one self-contained, human-readable file. It exists so a user reporting "the app feels slow to start" can attach exactly one file to a bug report, and any agent (or outside AI like ChatGPT) reading that pasted file can immediately tell which startup phase was slow without seeing the rest of the codebase. Test Regime Monitor A read-only developer-tools panel showing the health of your whole test setup in one place. Its Overview tab gives a single verdict — Healthy, Degraded, Breaking or Inconclusive — across six end-to-end stages, a System tab shows the machinery underneath, and a Branches tab shows each active branch's merge checks. It is hidden until you turn it on. UI Usage Tracking — see which controls and integrations you actually use Omniscio records on your own device which named controls you click and which sidebar integrations you open, then shows a used-versus-never-used report in Settings, with a Flow card for what you tend to do right after a control and an on/off toggle that stays separate from telemetry. Weekly Session Analysis (agent-friction dashboard + weekly deep analysis) A read-only developer panel that mines your past agent sessions for friction and waste. A free, always-on dashboard tallies timeouts, hook blocks, tool errors and redundant cd prefixes, plus spend concentration; a paid, opt-in deep run spawns one silent session that writes a ranked, lever-tagged report. Performance Abandoned helper processes are cleaned up When an agent's tool call gives up on a long command, the command itself keeps running with nobody waiting for it — and on a busy day dozens of these pile up holding memory and slowing the whole machine. Omniscio now finds those abandoned process trees, reports them first, and ends them once you turn enforcement on. Adaptive Memory Paging for Sessions An opt-in, Windows-only feature that caps the resident memory of every Claude CLI session when the host is short on physical RAM, so Windows pages the overflow out to disk. Sessions keep running, just slower. It is off by default, and it never kills a session to free memory. Agent action throttles What pacing Omniscio still applies to agent actions and what it deliberately no longer does: the retired filesystem and PID throttles that stay off, the correctness and safety protections that remain, and the authenticated admission broker that paces heavy work only when switched on. Box census (who is creating processes, and where the CPU actually went) The one command that answers "what is this computer doing right now": a ~10-second measurement of who is CREATING processes and which parent spawned them, how much CPU went to processes too short-lived to survive the window, the machine's fault and syscall counters, per-drive load, and per-process disk read and write. It keys every process on its identity rather than its name, and checks its own arithmetic before printing. Computer feels slow? A step-by-step fix-it guide (free fixes first, new hardware last) A plain-English, do-this-in-order guide for when your computer feels slow while running Omniscio — freezing for a few seconds, fans spinning up, everything lagging. It is written so you can follow it, and so an AI agent (any Omniscio session) can follow it for you. The steps go cheapest and easiest first. You almost never have to go far down the list. Buying new hardware is the very last resort — only after everything above it. Computer feels slow? A step-by-step fix-it guide (free fixes first, new hardware last) (part 2) Widen the start spacing (do this first if a burst is what hurts). A machine usually tips over when many things start in the same instant, not from the amount of work. Settings → Performance gives you two spacing dials — "Spacing between sessions when many restart at once" and "Wait between automation-created new sessions" — so a big wave trickles in instead of landing all at once. Same work, no spike. Full explanation under If RAM is fine but it freezes in bursts: a spawn storm below. Computer feels slow? A step-by-step fix-it guide (free fixes first, new hardware last) (part 3) (Developers only — skip if you don't use Omniscio to run coding sessions.) Git stores a project's history in two forms: fast, compact pack files, and slow, individual loose object files (one tiny file per saved object). Normally git tidies loose objects into packs on its own. Fast typing (isolated composer) An on-by-default setting that keeps typing smooth in very large sessions by giving the message box its own render path, so each keystroke stops redrawing the entire chat panel. It is a live kill switch — no restart — and every composer feature works the same in both modes. Frozen-panel retention (instant switch-back to recent sessions) Normally a session's chat panel is torn down when you switch away and rebuilt when you come back. Frozen-panel retention keeps the sessions you looked at most recently fully built in memory, so switching back is instant — and here is what it costs. Git storage compaction Git quietly accumulates junk until commands get slow and the store eats tens of gigabytes. The Git storage card lists every repository Omniscio knows about with its real size, then compacts the ones you pick — and it refuses rather than guess. Governance control and triage (what is being paced, and why) The one screen that says what the app is currently holding back and which governor decided it, and the one ordered runbook for working out why the machine is slow — both reading the app's own numbers rather than taking their own sample of the computer. They exist because three overlapping readouts used to answer the question differently. Heavy job pacing (retired) The old system that paced shell commands and queued background jobs is retired and off. What replaced it, what happens if the replacement is unavailable, and what the leftover Settings control now does — which is nothing. Idle session release Omniscio periodically releases the CLI process of a session that finished its turn and has been waiting on you past a threshold, reclaiming all of its memory. The conversation stays on screen; your next message brings it back transparently. Instant new session (Ctrl+T fast path) Starting a new session feels instant: the empty panel appears and the text box is focused in the same frame you press the key, with no spinner and no flicker. It works by quietly keeping one real, empty session prepared in the background, so the work that used to happen after the keypress has already happened before it. Isolated Session View (experimental) An opt-in performance mode that renders the active session's chat in its own background renderer process, embedded invisibly inside the main window. The app looks identical, but the conversation you are reading no longer shares a rendering thread with everything else, so switching into it and typing in it stay fast under heavy load. Keep chat responsive under heavy load (lean hidden panels) With many sessions open at once, every new chunk of output made all the background chat panels do bookkeeping — hundreds of tiny operations a second, enough to bog the app down and make chat scroll land in the wrong place. This stops the panels you cannot see and the visible one from reacting until you look at it. Keep my computer smooth under load (smoothLoadEnabled master toggle) "Keep my computer smooth under load" is a master on/off switch in Settings → Performance (on by default) for Omniscio's load-smoothing behaviors — the things Omniscio does so that running many Claude sessions at once doesn't bog your machine down. It is the umbrella control: turn it off and every load-smoothing behavior reverts to Omniscio's older, pre-smoothing behavior; turn it on (the default) and Omniscio quietly reduces the background work each session does. Keep Omniscio resident in memory (Windows working-set floor) A Windows-only Settings → Performance toggle that keeps Omniscio's own memory resident in physical RAM so its pages cannot be trimmed out to the pagefile, with a VirtualLock page pin as the primary mechanism, a per-process working-set floor as the fallback, and an option to serve the database from private memory. Keep recently-opened panels loaded (instant switching) Once you have opened one of Omniscio's heavier panels — the calendar, mail, Team Chat, the mind-map and whiteboard canvases — this keeps it alive in the background so returning to it is an instant flip instead of a cold rebuild. It is on by default, desktop only, and fully reversible. Lazy content load (instant cold-mount + truly lazy N-actions widget) Opening a session paints visible content within a few hundred milliseconds however long the conversation is, and the collapsible pill that hides a turn's tool calls now fetches its contents only when you open it. Together they shrink what a cold mount has to build by orders of magnitude on tool-heavy threads. Lite mode (optimize Omniscio for a less powerful computer) One switch that makes Omniscio lighter on a less powerful computer by flipping a curated bundle of Performance settings at once. It adds no new performance engine — it is a shortcut over knobs that were already there but scattered across about twenty toggles — and it is off by default and fully reversible. Low Power Mode A setting that trims the GPU-expensive visual effects — background blur, decorative hover shadows, the drifting blurred background — while keeping the clean, attractive parts. It slows busy animations and lightens re-renders, disables nothing functional, and takes effect immediately with no restart. Memory reclaim (free RAM automatically when the machine runs low) When your computer's free memory runs low and stays low, Omniscio frees some up so the machine does not start swapping to disk and feeling slow. Two levels of aggressiveness — a safe one on by default and a powerful one you opt into — triggered by a free-memory percentage you set. Omniscio priority boost + adaptive E-core delegation Windows-only Settings → Performance toggles that keep Omniscio's window and your other foreground apps snappy when many Claude sessions saturate the machine: boosting the app's own process priority, delegating sessions to efficient cores under load, and reserving a CPU core for the UI. Per-session update isolation (cheaper updates with many sessions open) An opt-in performance mode that makes every live update cheaper when many sessions are streaming at once: each open session panel reads its per-session state through one consolidated lookup instead of five, so a busy agent's output nudges far less work across every panel you have open. The app looks and behaves exactly the same. Performance Monitor (how the machine and the app are really doing) One answer to "how is this performing right now?", read three ways from the same measurement — a panel, a local route, and a command — so a panel reading and a scripted reading can never disagree. Every block says how much of the window it saw and which of four states it is in, so a missing number can never look like a healthy zero. Pre-render inbox sessions (instant switch into needs-you sessions) Omniscio builds the chat view of inbox sessions in the background, before you click, so opening one is instant instead of a brief loading shell. Because each built chat costs real memory, the work is bounded to the two hundred most recently active — the rest build on click, exactly like a running session. Runaway agent commands (the "has run N minutes" card) What the "An agent command has run N minutes…" inbox card means, what its Stop button does and never does, which commands Omniscio stops on its own, and how to turn the watch off. One runaway command can slow every agent on the machine; the watch finds it and lets you end just that command. Session CPU cap (Windows soft-cap + below-normal priority) An opt-in, Windows-only cap that bounds the CPU the whole Omniscio process tree may use once other programs are competing for the machine, alongside the below-normal priority that already makes sessions yield. It covers sessions, the subagents they spawn, tests, builds and MCP servers alike. System file cache cap (Windows) — stop the file-cache freeze What it is. A Windows-only control in Settings → Performance (and a one-click button on a specific inbox alert) that installs an opt-in hard cap on how much memory Windows uses as its system file cache. It exists to prevent a specific whole-computer freeze. The freeze it prevents. When many things read and write files fast — for example dozens of Omniscio agent worktrees churning on a Windows "Dev Drive" (ReFS) — Windows can balloon its file cache until it eats most of your RAM. The load curve (what agent commands cost at different session counts) A readout that buckets how long agent commands actually wait by how many sessions were live at the time, joined against the machine's own stall and governor data. It ends by naming the command class that degrades fastest with load — and, since 2026-09-22, by putting what the fleet FINISHED in each bucket beside the waits — and never by recommending you run fewer sessions. Tune background pacing during severe lag Omniscio can slow background agent starts when it detects severe interface lag. You can adjust that protection while the app is running; no restart is required. Open Settings → Performance → Sessions, expand the advanced controls, and change Seconds between background starts during severe lag. 10 seconds is the recommended default. Higher values protect responsiveness more aggressively by starting background work less often. Work output by load (what the fleet actually finished) The readout that answers "what did the fleet actually FINISH at this load?" — commits landed, ready-to-merge tags applied, and items brought to the owner's inbox, each per hour for the whole fleet and per session-hour. Tokens are a supporting column, because they measure how much was SAID, not what was DONE. Beneath it sit two more axes: a SESSION LIFECYCLE section, and a SESSION TIME section naming where every second of a session's clock went, with the leftover reported as UNACCOUNTED. Something's not working? Start here The front door for when Omniscio is misbehaving. Find the line that matches what you're actually seeing, and follow it. Every destination is a real page with the full fix — this page only points. You don't need to know what caused it. Describe the symptom to yourself in plain words and pick the closest match. --- CLI & API Reference 22 Driving Omniscio from outside — the local control server, its routes and look-up tables.
CLI Control Agent Tools (consolidated dashboard) Agent Tools is the sidebar group that nests Omniscio's tool and key views under one collapsible row, plus a dashboard panel that counts CLI Tools, Skills and MCP Servers, with API Keys, Agent Email and Browser Logins nested below them. CLI Control (control Omniscio from scripts and hotkeys) CLI Control is the local-only HTTP door into Omniscio: how to turn it on, copy or regenerate its token, drive the app from a script or a hotkey, and what it refuses to do without your say-so. This overview covers what the server is, where its switch and token live, and the walkthrough for calling it. CLI Control (control Omniscio from scripts and hotkeys) (part 2) Part 2 of the CLI Control page: the endpoints an outside script or AI uses to change things on your behalf — driving a running session, registering, editing or deleting a project, adding and removing project docs, away-mode rules, and firing a recipe — plus how the approval queue and the rate caps treat each one. CLI Control (control Omniscio from scripts and hotkeys) (part 3) Part 3 of the CLI Control page: the surfaces an outside AI can look at and drive — AI Coaching interviews and artifacts, bookmarks, inbox alerts, a snapshot or screenshot of the window, an on-screen spotlight it can point with, and a jump straight to one setting. CLI Pending Actions (approve external AI changes) When an external AI asks Omniscio to change a setting, alter a session, or take another guarded action over the CLI control server, the request lands in your inbox as an approval card instead of applying itself — what that queue is, how to choose which action families need your click, and how to read and answer a card. CLI Pending Actions (approve external AI changes) (part 2) Part 2 of the CLI Pending Actions page: what happens after you decide — how the session that asked is told your answer, what the queue does when a card is never answered, which settings can never be patched from outside, the note rules, the in-app bypass, and the ready-mint override card. Omniscio Control (how an AI agent changes the app live) A bundled skill that teaches an AI agent how to read and change a running Omniscio install through the app's own control server — settings, sessions, scheduled work, the inbox — instead of editing files on disk. Agent-facing: you just ask in plain English, and the skill carries the rules, the approval model and the error handling. Recent CLI Activity — an audit log of what scripts and agents do through Omniscio's local server A read-only audit card in Settings that lists the calls made to Omniscio's local control server — one row each, in plain English, with the outcome, a coarse idea of who called it, and the session it spawned or relates to. Covers what is recorded, what is deliberately skipped, how long each kind is kept, and why it is a breadcrumb rather than a security control. Tools View (browse installed CLIs + linked skills) The Tools view is a dedicated virtual project that lists every CLI Omniscio integrates with, across ten categories, with each row showing install status and its detected version. Selecting a tool opens a detail pane with the description, an Install button for auto-installable tools, and chips for the skills that depend on it, which jump to that skill in the Skills view. API & Routes Account re-login over HTTP — `POST /accounts/login-capture` The HTTP twin of the account panel's general Log In button: how an agent starts an Anthropic account re-login through Omniscio's local control server, polls until the flow finishes, and confirms which account was captured. Covers the two routes, the full headless sequence, every error status, and the credential-safety gates wrapped around it. Agent-Driven Sessions (External AI Drives an Omniscio Session) An external AI can ask Omniscio for a brand-new session over HTTP, get one approval through the inbox, then drive that session through bounded multi-turn dialog under turn and dollar caps, with a revoke control always available in the session header. AI Coaching via CLI (manage coaching artifacts and start interviews from an AI agent) AI Coaching via CLI lets an outside AI agent browse the interview-prompt library, list and read saved coaching artifacts and their version history, propose new artifacts, edits and deletions, and start a new interview — over Omniscio's local control server. Covers which routes apply immediately, which wait for inbox approval, and idempotency. Change keyboard shortcuts by asking AI Any Claude Code session on your machine can read and rebind Omniscio's keyboard shortcuts in plain English once CLI Control is enabled — no clicking through Settings, and the change takes effect immediately with no approval step. Covers the phrasings that trigger it, the validation rules a binding must pass, and the conflict report. CLI Feedback Submission (file bugs / feature requests from external AI) How to file a bug, feature request, feedback or question from an outside AI: the in-app Send Feedback button and the local HTTP route that mirrors it, what each response code means, how diagnostics get collected, and the rules an agent should follow so a report is actually useful to the developer reading it. CLI Session Recovery — Get Stuck Sessions Going Again Three targeted CLI calls let an AI agent diagnose and fix a wedged Claude session with no approval round-trip: moving a session off an account that hit a rate limit or auth error, restarting a process that died, and nudging a session that is alive but idle. Covers finding a healthy account and what each call returns when it goes wrong. CLI Tags Control (manage the tag library from external AI) The eight local HTTP routes that let an outside AI agent or script list, propose, edit, and soft-delete your tag library, and apply or remove tags on individual sessions — including which routes need your approval in the inbox and which apply immediately. Deep Links (omniscio:// URLs) Omniscio's custom link protocol: the omniscio:// URLs that open the app and jump straight to one inbox approval, project, session, note, mind map, setting, Supermail thread or team-chat message, plus the older agentmc:// form still accepted on the way in. Covers every route, how to build the links, and what the same links do on a phone. Deep Links (omniscio:// URLs) (part 2) How deep links are implemented: the resolver behind them, what happens when an agent writes a link that does not resolve, the route each link takes into the app's navigation, and the contracts and tests that hold the behaviour in place. Quick Replies via CLI (read, create, move, delete, export and import the library from an AI agent) Quick Replies via CLI exposes the personal Quick Replies library — snippets, dividers and folders — to an outside AI agent over Omniscio's local control server: ten routes to read, create, edit, reparent, reorder, delete, restore, export the whole library as a portable bundle, and import one back. Covers the gating model, the flat and tree read shapes, the add-only import merge, and what is deliberately out of scope. Share via CLI (publish, list, revoke, delete from a script or AI) Four endpoints on the local CLI control server at http://127.0.0.1:19519 let any tool that can make an HTTP request publish files or pasted content as shareable links, list existing shares, revoke (soft-disable) a share, or delete (hard-remove) a share row. SMS via CLI (read, triage, and send texts from an AI agent) Reading, triaging and sending texts from outside the app, over Omniscio's local control server, so an AI agent or a shell script can work your SMS inbox. Covers the three tiers the routes are split into, the workflow a caller follows, the approval gate that stands in front of sending, and what is deliberately left out. Reference Tables Legal & Policies 4 Terms, privacy and the other policy pages.
Privacy & Data Auto-delete old message content (opt-in communications retention) The opt-in setting that permanently deletes old communications content — messages, email samples, voiceprint data, meeting recordings and transcripts, and the digests built from them — after a window you choose, plus the one-time notice and 30-day grace period that guard it. Database Encryption Opt-in encryption of your live local database at rest: the two tiers (automatic OS-keychain unlock and a zero-knowledge passphrase), the mandatory recovery code, what happens at the next restart when you enable or disable it, and the guarantees that never regress. Is my data encrypted? A map of what Omniscio encrypts and what it does not, so the exact answer is one row away. The copies it makes of your data and your stored secrets are encrypted; the live database is not, by default — and this page covers the setting that changes that, and what each surface relies on instead. Secret Paste Guard Secret Paste Guard is a shipped, opt-in feature that keeps API keys and other secrets out of your AI conversation history: paste a secret into the session composer and Omniscio stores it locally under an opaque handle, substituting the real value only at execution time on Claude-based engines. Uncategorised 22 Pages not yet filed under a product area.
Agent conversations Agent conversations AI, accounts & models settings AI, accounts & models settings AI-Drivable Browser (in development — enable it in Settings → Lab) The AI-drivable browser — in development, and live again behind its Settings → Lab toggle (`aiBrowserEnabled`): it turns the embedded Browser pane multi-tab and lets an Omniscio agent drive it in its own isolated tab, with an autofill vault, Memories, Journeys, Design mode, workflow recording and CLI routes. For driving your own signed-in Chrome, use My Real Chrome instead. AI-Drivable Browser (in development — enable it in Settings → Lab) (part 2) The second half of the AI-drivable browser page: what the feature added over the plain embedded browser, Design mode and Visual Edit, spaces and quick navigation, and the agent's own isolated tab with its MCP tools and permission tiers. Appearance & display settings Appearance & display settings Auto-lander & dev-pipeline settings Auto-lander & dev-pipeline settings Cashbox (side-project finance plugin) Cashbox is a first-party plugin that tracks money in, money out and profit for each of your side projects. You type what happened in plain words ("paid 20 for hosting yesterday") and it saves the entry at once, reads it with your AI account, converts other currencies at the European Central Bank rate for that day, spots repeating costs and exports everything to a spreadsheet. It ships switched off, keeps your numbers on this computer, works with the AI switched off, and lets other Omniscio sessions log entries but never delete them. Controls, permissions, privacy & general settings Controls, permissions, privacy & general settings Integrations & connections settings Integrations & connections settings More settings (A) More settings (A) More settings (B–G) More settings (B–G) More settings (H–Q) More settings (H–Q) More settings (R) More settings (R) More settings (S) More settings (S) More settings (T–Z) More settings (T–Z) Notifications, inbox & voice settings Notifications, inbox & voice settings Omniscio — Feature Recommendations (by goal) Omniscio — Feature Recommendations (by goal) Performance settings Performance settings Review tiers (how much review a change gets) The session asking for an AI review says how much review the change needs — light, standard or heavy — and Omniscio runs the reviewers you configured for that size. Any change touching a database migration, signing in, or money, or one simply too large to read closely, is raised to the heavy line-up whatever was asked for, and the app records what it asked for, what it ran, and what the change looked like. You configure each size, and an optional trial AI, in Dev Pipeline → Setup. Sessions & agents settings Sessions & agents settings Settings documentation The one entry point to Omniscio in-app settings help — the per-setting pages behind the small ? button on every Settings screen, grouped by area. Every page here is offline and opens in place, right next to the setting it explains. Settings reference (every configurable setting) Every setting Omniscio can store, one row each: its key, the kind of value it takes, the values it accepts, and what it does. Use it to answer "is there a setting for X?" and "how do I configure X?" without opening the app. > Generated from the settings schema by `npm run settings:catalog` -- do NOT edit by hand. It is rebuilt whenever the app changes, so it is always current. Each row is one setting: its key (the id used by `PATCH /settings/:key` and shown in `GET /settings`), the kind of value it takes, the accepted values, and what it does. Rows are grouped by area below, and each area is split into the toggles and fields shown in the Settings UI and the CLI/advanced keys that have no Settings control. Secret credentials and server-managed keys are omitted (they are not user-settable).