# More settings (S)

Reference entries for all Omniscio settings whose key begins with the letter S.

## Scheduled Messages {#scheduledMessagesEnabled}

Turns the Scheduled Messages virtual project on or off. When enabled, a "Scheduled Messages" entry appears in the sidebar where you can compose a rich-markdown note to yourself and have it dropped into your inbox once or on a recurring schedule — a simple way to set future reminders without creating a full session. On by default. The feature ships on; what starts hidden is its sidebar row, which is a separate toggle (`scheduledMessagesSidebarEnabled`) — the feature and its inbox delivery still work with the row hidden. Turn either one off in Settings → Features.

## Screen Recorder {#screenRecorderEnabled}

Shows the Screen Recording entry in the sidebar. Once enabled you can record your screen (with webcam overlay, multi-source layouts, and global hotkeys), then edit clips and browse them in a built-in library. While enabled, the recorder controls and configured countdown stay loaded so Quick Record can respond immediately; no screen, microphone, or camera is captured while idle. Recording quality, frame rate, click-ring style, watermark, countdown, and dozens of other capture options live inside the Screen Recording panel's gear menu, not in a separate Settings page. Off by default; turn it on in Settings → AI Coach.

## Screenshots {#screenshotsEnabled}

Shows the Screenshots entry in the sidebar — a built-in snip tool paired with its own library. Snip any region of your screen with a global hotkey (Ctrl/Cmd+Alt+S), then annotate, copy, or share the result. Turned off by default for new installs; users who previously had "Screen Capture" enabled keep their Screenshots area. When off, the sidebar row is hidden and the snip hotkey stops working. Turn it on in Settings → AI Coach.

## Daily restart budget per session {#sessionRefillMaxRevivesPerSessionPerDay}

How many times the session refill governor may restart any one session in a single day. The cooldown is only a delay between attempts; this is the budget, and it is what stops the governor circling back to the same stuck session over and over. Leave it at 1 unless you specifically want repeat nudges. Range: 1–50.

## Session title style {#sessionTitleStyle}

How AI-written session titles are phrased: **brief** (a few words), **descriptive** (the default — a short summary of the work), **journey** (how the session has evolved), **latest topic** (what it is about right now), or **custom**, which uses the prompt you write yourself. Only consulted when session naming is set to AI titles.

## Cross-post share comments to Team Chat {#shareCommentsCrosspostTeamChat}

Controls whether comments left on a shared artifact are automatically posted into a linked Team Chat channel. Default on. Turn it off if you want share feedback to stay in the Shares inbox only, without making noise in a channel.

## Enable commenting on new shares by default {#shareCommentsDefaultOn}

When on (the default), every share you publish starts with commenting enabled, so viewers can leave feedback right away. Turn it off if you prefer shares to be view-only unless you explicitly enable comments for a specific share.

## Inbox notifications for share comments {#shareCommentsNotifyInbox}

When on (the default), a new comment on any of your shared artifacts raises an inbox notification so you don't miss feedback. Turn it off to stop share-comment notifications while keeping all other inbox behavior unchanged.

## Open new shares automatically {#sharesAutoOpenOnCreate}

When on, a share you publish opens immediately in the built-in viewer. A share an agent publishes shows a one-click "Go there" toast rather than forcing a screen switch. Default off (opt-in). Has no effect unless "Open share links in app" (`sharesOpenInApp`) is also on.

## Open share links in app {#sharesOpenInApp}

When on (the default), clicking a share link opens it in Omniscio's built-in viewer rather than your system browser. The built-in viewer lives in the main window only, so a share link clicked in a popped-out session or project window, or in the KMS, Writer or project-doc windows, opens in your system browser instead. Desktop only — on web or mobile the link always opens in the browser regardless of this setting. Accepted values: `true` (in-app) or `false` (system browser).

## Tell agents the shell remembers its folder {#shellCwdPersistNudgeEnabled}

On by default. Adds a short note when a session starts telling the agent that its shell keeps the folder it changed into, so it stops re-entering the same folder before every single command. It removes a small tax that every default-settings user otherwise pays on every shell command, which is why it ships on. Purely an instruction to the agent; it changes nothing about how commands actually run. Turn it off to drop the note. Settings → Features.

---

## Show alternative AI providers {#showAlternateProviders}

Reveals the additional AI provider panels (Codex, Gemini, Anti-Gravity, DeepSeek, Kimi, OpenCode, Cursor) in the new-session menu and in Accounts & Providers. Off by default — Omniscio uses Claude only until you turn this on. You can enable it from Settings → Accounts & Providers.

## Show bookmarks toolbar button {#showBookmarksInToolbar}

Shows the bookmarks button in the header toolbar, which opens a quick-access popover of your saved URLs and links. On by default. Turn it off in Settings → Widgets to hide the icon; the popover becomes unreachable, but your saved bookmark rows are preserved and reappear if you turn it back on.

## Show recipe pipeline lane sessions {#showRecipeLaneSessions}

Reveals the per-lane sub-sessions that recipe pipelines spawn in the sidebar. Off by default to keep the sidebar uncluttered — lanes auto-archive on clean completion anyway. Notifications remain silent for lane sessions regardless of this setting. Turn it on if you want to watch individual pipeline stages as they run.

## Show context-use indicator {#showSessionContextIndicator}

When on, a context-percent ring appears in the session header so you can see how full the context window is at a glance. Hidden by default to keep the header clean. Turn it on in Settings → Sessions if you want a quick visual check before a session runs out of context.

## Show subagent activity {#showSubagentActivity}

When on, a chip appears in the session UI while the agent has dispatched subagents (Task tool fan-outs) that have been running for more than about 15 seconds, and a collapsible result card is added to the transcript when each subagent finishes. The same cards also carry a subagent's own narration as it works, attributed to the subagent — that text is never folded into the session's reply, so a subagent's report can never render as the session's final message. Off by default so the extra detail stays opt-in. The underlying tracking that keeps the stall watchdog patient during long fan-outs runs unconditionally.

## Sidebar position {#sidebarPosition}

Places the project and session sidebars on the left or right side of the window. Accepted values: `"left"` (default) or `"right"`. Change it from Settings → Appearance or from the Settings home page's popular shortcuts.

## Silence recipe steps on clean completion {#silenceRecipeStepsOnSuccess}

When on, mainline recipe step sessions and the orchestrator are automatically archived when they finish cleanly — no errors, questions, or trouble markers. Off by default, which keeps completed recipe sessions visible on your board. Turn it on to declutter the inbox after long audit or pipeline runs.

## Skill slash-command autocomplete {#skillSlashAutocomplete}

When on (the default), typing a leading "/" at the start of the message composer opens an inline menu listing your installed skills — filtered as you type — so you can insert a skill command with arrow keys, Enter/Tab, or a click. Only triggers when "/" is the very first character, matching when the Claude CLI treats input as a slash command, so a mid-sentence slash (like "and/or" or a file path) never opens the menu.

## Personal Slack account {#slackPersonalEnabled}

When on, Omniscio sends Slack messages as you (OAuth, your personal account) rather than as the Slack bot. Requires bot setup to be completed first. Turn it on in the Slack settings after connecting the bot if you want outbound messages to show your own identity.

## Snooze notes {#snoozeNoteEnabled}

When on, the snooze palette shows an optional note field so you can leave a short reminder for yourself that appears when the snooze fires — for example, "follow up on the invoice" or "check if the build passed." Off by default.

## Soniox voice ID {#sonioxVoiceId}

The voice ID used when Soniox is selected as your text-to-speech provider. Enter a preset voice ID from the picker in Settings → Voice, or paste a custom Soniox voice ID into the text field to use a voice not listed in the presets. Empty means no voice is selected for Soniox.

## Speechify voice ID {#speechifyVoiceId}

The voice ID used when Speechify is selected as your text-to-speech provider. Pick from the available Speechify voices in Settings → Voice, or enter a custom voice ID manually. Requires a connected Speechify API key.

## Sticky Notes {#stickyNotesEnabled}

When on, enables floating sticky notes on your desktop — small always-on-top windows you can pin anywhere on screen for quick references or reminders. On by default; turn it off in Settings → Features.

## Auto-summarize mail from these senders {#supermailAiSummaryAutoSenders}

The senders whose emails get a "Catch me up" summary automatically the moment you open them, without you asking. Each entry is either a full address or a bare domain like `acme.com`, which matches any address at exactly that domain (never a subdomain, never a partial match). Empty is the default, so nothing is summarized automatically until you add someone, and every summary is a deliberate click until then. Manage the list in Settings → Email & Summaries, or toggle a sender straight from an open email. Holds up to 200 entries.

---

## Supermail "Catch me up" AI model {#supermailAiSummaryModel}

The AI model that writes the Supermail "Catch me up" thread summary. Empty (the default) uses the built-in cheap OpenRouter model with Claude Haiku as a fallback. Pick any model from the catalog to override — its provider is resolved in the main process and Claude Haiku is always kept as a backup for unavailable picks. Choose it from the model picker in Settings → Email & Summaries. Only consulted when the AI summary feature is on.

## Supermail "Catch me up" prompt {#supermailAiSummaryPrompt}

A custom prompt (up to 4,000 characters) that replaces or supplements the default instructions given to the AI when it generates a Supermail thread summary. Leave it empty to use the built-in prompt. Only consulted when the AI summary feature (`supermailAiSummaryEnabled`) is on.

## Supermail "Catch me up" trigger {#supermailAiSummaryTrigger}

Controls when the Supermail AI thread summary runs. `"manual"` — you click a "Catch me up" button to trigger it. `"auto"` — the summary runs automatically each time you open a thread. Default `"manual"`. Only applies when the AI summary feature is on.

## Supermail row spacing {#supermailDensity}

How much vertical space appears between emails in the Supermail list. Accepted values: `"compact"`, `"comfortable"`, or `"spacious"`. Adjust it in Settings → Email & Summaries to match your preference for inbox density.

## Show emails in Omniscio inbox {#supermailEmailsInInbox}

When on, each incoming Supermail email appears as a card in the Omniscio inbox alongside session notifications, with an inline reply box so you can respond without switching to the Supermail panel. Turn it off to keep Supermail mail out of the main inbox.

## Supermail (sidebar) {#supermailEnabled}

Master switch for the Supermail panel. When on, the Supermail entry appears in the sidebar — a full email client (folders, compose, snooze) that connects to Gmail or IMAP. Off by default while the native port is in preview; turn it on in Settings → AI Coach.

## Supermail header layout {#supermailHeaderLayout}

How Supermail arranges the Mail/Sessions switch and the mail toolbar in its header. Accepted values: `"current"`, `"unified"` (default), or `"rail"`. A personal display preference with no impact on mail data.

## Supermail sidebar badge count {#supermailInboxBadgeMode}

What the number next to Supermail in the sidebar counts. `"unread"` — drops as you read messages (like Apple Mail or Gmail). `"total"` — drops as you archive, snooze, or label (like Superhuman's inbox-zero count). Default `"unread"`.

## Supermail email summary model {#supermailSummaryModel}

The model used for Supermail email summaries (video summaries and archive digests). "Automatic" (the default) inherits the global Utility AI model. Pick a specific model from the catalog to override it for Supermail summaries only. Find it in Settings → Sessions → Per-feature AI models.

## Use connected Google account for Supermail {#supermailUseAmcGmail}

When on, Supermail reads your inbox through the Google account you already connected in Settings, so no separate Supermail sign-in is needed. Turn it off to use Supermail's own independent Google sign-in instead.

## Who pays & who serves {#supplyLists}

For each AI model family you can run besides Claude (DeepSeek, GLM, Kimi, MiniMax, Meta, Qwen and OpenRouter), an ordered list of rows saying who pays for a session and which company serves it. A row is your own account with a company (the model's maker, or a reseller such as DeepInfra), Omniscio credits (Omniscio picks the company for its own credits), or the company lane for Omniscio developers. When a session connects it goes through the first row that can serve right now: a row with no key, no sign-in or no credit is skipped, and when a row runs out of credit its account is skipped by every session (in every family that pays from it) until Omniscio tries it again — one session every 10 minutes — or a balance check sees money. The running session moves to the next row and carries on, and one inbox card per account says so. The model you picked never changes, only who pays and which company answers. A reseller row whose company gives no discount on repeated (cached) text is marked, because the same work can cost several times more there; DeepInfra gives the discount, so its DeepSeek rows carry no mark. A DeepSeek row can be limited to DeepSeek’s peak-price hours (the moon button, or `"onlyDuring": "deepseek-peak"`), so it serves only while DeepSeek charges double. Cloud sessions use your own account with the model's maker, or Omniscio credits.

Edit it in Settings → Accounts → Who pays & who serves, as rows or as JSON text. The lists start from the keys and choices you already had, and a family you leave out of the JSON gets its starting list back; an empty list means that family's sessions cannot connect. Stored as an object mapping each family to its rows, for example `{"deepseek": [{"pay": "you", "vendor": "deepseek"}, {"pay": "omniscio"}]}`: `pay` is `you`, `omniscio` or `company`, and `vendor` names the company on your own account.