Omniscio

Omniscio — Subprocessors

Effective date: 2026-08-20 Last updated: 2026-11-17


Key terms at a glance

Plain-English summary — not a substitute for the full Subprocessors list. This box is a quick, friendly overview of the points people ask about most. It is not legal language and does not replace the sections below, which are what actually applies. If anything here seems to differ from those sections, the full sections control.


What this page is

A subprocessor is a third party that processes personal data on Omniscio's behalf, under Omniscio LLC's own accounts, to help us provide the Services. This page lists those subprocessors, what each does, and the data involved.

It does not list the many services you can connect on your own account or API key (for example, an AI provider you bring your own key for, or a Slack or Notion workspace you link). For those, you deal with that provider directly under its own terms and we are not acting as your processor — see "Services you connect on your own account" below.

The Services are currently offered only in the United States, and the data we store is stored in the United States (see Privacy Policy §4) — with one exception, shared video recordings, which are stored in Cloudflare R2's Eastern North America (ENAM) region (Cloudflare's placement for that bucket, not a residency guarantee, so we do not promise United States storage for them — see "Where processing happens"). Some of the vendors below are not US-based — see "Where processing happens" for the specific list. Your data is handled in the US, except where you choose a model or service hosted outside the US, which is your choice. We are putting data-processing agreements (DPAs) in place with the subprocessors below; a DPA is available for business/team customers on request at legal@omniscio.com.

Where processing happens

Our own storage and backend are in the United States — accounts, database, files, team and Shares content, and error reports all live in a US region. The one exception is shared video recordings: the mp4 of a video Share is stored in Cloudflare R2 (the Cloudflare row under Core subprocessors), in Cloudflare R2's Eastern North America (ENAM) region. That region is Cloudflare's placement for the bucket rather than a contractual commitment, so we do not promise that a shared recording is stored in the United States.

Being offered only in the US bounds where our customers are. It does not bound where a vendor is. These vendors process content outside the United States:

Vendor Where What it processes When
DeepSeek China The prompt/message content of the coding lanes routed to it on a Company key (served first-party by DeepSeek, failing over to DeepInfra for most models). Never Google user data, and never a built-in helper feature. Only while no Google account is connected
CrofAI China The prompt/message content of a coding lane routed to crofai/* on a Company key (a Pro, tier-gated lane — retired 2026-09-21, existing sessions only) When you choose Omniscio credits for the CrofAI lane
Zhipu (GLM) China The prompt/message content of a GLM coding session routed to Zhipu (api.z.ai) on a Company key Only when you choose Omniscio credits for a GLM session (Pro)
OpenRouter → Moonshot (Kimi) / MiniMax China (onward leg) The prompt/message content of a coding lane routed to Kimi or MiniMax on the pooled Company OpenRouter key When you choose Omniscio credits for that vendor
Fish Audio China-founded, no region stated The response text read aloud, forwarded verbatim through our relay Only if you turn on voice (off by default)
OpenRouter → Alibaba China (onward leg) Some OpenRouter helper requests are served by an Alibaba-operated model upstream When our gateway routes a helper request there
OpenRouter → SiliconFlow China (onward leg) The request text of an Instant Fix run on the Gemma 4 model, which OpenRouter can hand to SiliconFlow among its hosts Only when you choose Gemma 4 for Instant Fix (the default model is a different one)
Google Analytics Global (Google) Your IP address, browser user-agent, the pages you view, and a randomly-generated analytics identifier — on the pages of our public websites, these legal pages included (the help library, the /welcome page and the account-email link page carry none), and loaded only while your consent choice permits analytics. Google sets its own _ga cookies, and its advertising features ("Google signals") are on, which our Privacy Policy treats as a "share" and, in some states, a "sale". Consent-gated: no visitor is opted in by default — analytics and ad storage stay denied until you accept the banner, in every time zone — and you can opt out at any time from the "Do Not Sell or Share My Personal Information" link in our footer or with Global Privacy Control. Every visit to a page on this website, including this one — once you accept
Meta (Meta Pixel and Conversions API) and Google Ads Global (Meta Platforms, Inc.; Google) Only while we run ads on that platform, and only for a visitor with advertising consent: the Meta Pixel and the Google Ads conversion tag load after that consent — never on a first download click — sending your IP address, user-agent and Meta's _fbp / _fbc / fr ids, and — if you then create an account — a signup and purchase event sent from our server to Meta, carrying a one-way hash of your account id and email, _fbp / _fbc, your user-agent and, for a purchase, its amount. Both process it under their own terms as independent third parties. A "share" under California law — opt out with Decline, the "Do Not Sell or Share My Personal Information" link, or Global Privacy Control. See our Cookie Notice. Only while we run ads on that platform, and only after you consent to advertising — a signup / purchase after a download

Some of the services you connect on your own account are also non-US — Moonshot/Kimi, Zhipu/GLM and MiniMax are China-hosted, and the OpenStreetMap Nominatim geocoder is EU-hosted. When you connect them on your own account, those run under your account with that provider, so the relationship (and the transfer) is directly between you and them. But the Company also runs lanes for some of them on Company-held keys — the rows above for GLM (Zhipu), Kimi / MiniMax, DeepSeek and CrofAI — so on those lanes the relationship and the transfer are ours, not yours; they apply when you choose Omniscio credits. CrofAI's Company lane is retired (2026-09-21), so no new session can reach it.

Core subprocessors (used to run the Services)

Subprocessor What it does for us Data it may process
Anthropic (Claude) Default AI provider for agent sessions. If you set up an AI bot that answers in a Team Chat channel, the messages it answers also become that session's prompt — including other members' messages. The optional personal-development features (AI-coaching, journal reflection) also call Anthropic to generate insights — normally on your own Anthropic API key, though a key already stored in the App (including a Company-held key on a plan that includes AI credits) can fund those calls instead. Mission Control's What Next also asks Anthropic for Claude through our gateway, on your own Anthropic key or on a Company-held one — its answer, the lookups behind it, and its deeper look, which uses a more capable Claude model Your prompts, code, file contents, conversation context, and attachments; and, for the personal-development features, the coaching/journal text those insight calls process (on your own key, or on a stored / Company-held key when one funds the call). What Next sends the question, the board context, and the matching snippets from each source that person has connected and allowed — email, calendar, Team Chat, board items, GitHub, Jira, Linear, Asana, ClickUp, Notion, OneDrive, Dropbox, plugins, Vault notes, Google Docs and meetings. For a Team Chat AI bot, the triggering message, the surrounding channel messages and the attachment file names are sent; member display names are replaced with pseudonyms and guests are excluded, but the message text itself and the file names are not removed
Google — Firebase / Google Cloud Authentication, database (Firestore), cloud functions, hosting, our AI gateway (Cloud Run), secret storage, mobile push (FCM, which hands Apple devices' notifications to Apple's push service), live Gemini speech-to-text for dictation when you choose that engine and have no Google key of your own, the encrypted sync locker (Cloud Storage, US) — only if you turn on Cross-Device Sync — and the encrypted cloud backup you upload to Omniscio (Cloud Storage) — only when you start one — plus the cloud test runner's storage bucket, which holds the code bundle and screenshots uploaded for a run and that run's logs and result reports, not end-to-end encrypted and bot protection (reCAPTCHA, through Firebase App Check) on the public pages of the Forms plugin Account/identity, billing and metering, usage/spend aggregates, support content, diagnostics, and team/Shares data. Cross-Device Sync: the end-to-end encrypted sync payload, which we hold as ciphertext we cannot read, plus the small plaintext device registry (per-device id, the label you choose, the device's public key, and timestamps). Cloud backup: the backup archive, encrypted on your device before upload (its key never leaves your computer), kept until you replace or delete it. Cloud test runner: the source bundle a --cloud run uploads (your own working-tree files, plaintext in a bucket only our service identities and your own account can read), the run's captured logs and result reports, the screenshots a run captures, and the machine identifier and session id recorded against each submitted job so the run can be attributed and erased on request; retained on a fixed bucket lifecycle and deletable per machine or session through our erase command. Gemini dictation: the live microphone audio you dictate. Forms bot protection: the visitor's IP address and browser signals
Stripe, Inc. Payment processing and US sales-tax calculation Payment method and billing details (held by Stripe); we receive a customer identifier and subscription status
Sentry Application crash/error reporting Crash/error reports with scrubbed messages
Resend Transactional email (bug reports, digests, notifications) and product-activation / onboarding email Email content the App sends; for activation/marketing messages, your email address and the message content
SMTP2GO Transactional email sent by our self-hosted backend apps — account invitations and password resets The recipient's email address, the message subject and body, and any attachment. That recipient can be an invitee you invited, who is not a user of the Services at all.
Twilio SendGrid Transactional email for the Calendar Scheduler booking feature — confirmations, approval requests, reminders, and cancellation / reschedule notices The invitee's email address, their name, and the booking details carried in the message, including anything they typed into the notes or the host's own questions.
Cloudflare Agent-email relay (inbound and outbound) and its outbound send leg; account security / deletion / tier notices through Cloudflare's Email Sending API; storage of shared video recordings in Cloudflare R2; bot protection (Turnstile) on public forms Email routed through the agent-email feature, including sender/recipient, subject, and body; the recipient and content of the account notices; the mp4 of every video Share — screen content plus microphone and system audio, and, when the camera overlay is used, the webcam video of you and anyone in frame composited into that same mp4 (held in Cloudflare R2's Eastern North America (ENAM) region — a placement Cloudflare applies to the bucket, not a residency guarantee); and, on a public form, the visitor's IP address and browser signals
OpenRouter (via our AI gateway, on Company-held keys; OpenAI models are reached through it) Cost-efficient "helper" AI models for built-in features — none of which uses DeepSeek — inbound agent-email pre-screening/classification (OpenRouter, on google/gemini-2.5-flash-lite), session titles, summaries and digests, Plain Speak rewrites, contact-name inference, Team Chat AI (a channel summary built from up to that channel's newest 100 messages, and @mc answers from the boards a channel's members share — the summary carrying other channel members' message text, with each author appearing only as a pseudonym such as Participant 1 rather than their display name), email filtering/thread summaries, and Inbox Pilot (classifying the recent turns of an agent session, and drafting a reply when you enable that), and Tasks AI planning — ranking what to do right now, the daily check-in and its chat, Plan my day, the Break it down interview, the arrange pass, Plan chat, the session-end learnings distill, and the session-end catch scan (watching finished sessions for follow-up tasks), and Mission Control AI helpers and Omni — the advisor's quick question, sprint retrospectives, import column matching, the activity summary, the busy-season and risk check-up, and Omni reading a document, a spreadsheet or a form's answers; OpenRouter also serves the off-by-default Voice conversation ("Layer 2") Lab feature, where each spoken turn is decided on a Grok (xAI) model from the open sessions' names/status/summaries, the active session's transcript tail and session-search snippets; and OpenRouter serves the last fallback of the Ask Omniscio DeepSeek lane (a DeepSeek model — see the DeepSeek row) The prompt/message content those helper features process — Tasks AI planning asks for the OpenRouter helper lane but is not guaranteed to reach it: the App's key order decides which key pays and which vendor receives the request — an AI key you have saved, then Omniscio credits, then the Company allowance this row describes. It sends your task text, task-list context, About You planning profile, Tasks memory/learnings, check-in and Plan chat messages, and — from the session-end catch scan — up to about the last 12,000 characters of the session it reads (your turns and the assistant's). Plan my day also carries your day's calendar: meeting start and end times, plus event titles when Show meeting names in AI planning is on. including the daily digest (on by default), which carries the first message of every session it summarises — AI-coaching and journal sessions are excluded and never appear in it. Team Chat channel AI is not guaranteed to reach this lane either — it follows the same key order, and it sends the channel transcript it summarises (up to that channel's newest 100 messages, each line prefixed with the sender's pseudonym — Participant 1, Participant 2 — never their display name) and, for an @mc answer, the asking member's question plus the shared board context. Mission Control AI helpers send board item names, comments, the names of the people on a board, form respondents' answers, up to 20,000 characters of a document, the first rows of a spreadsheet, the column names and a few sample values from a file that person imports, and up to 100,000 characters of the document a person hands to the busy-season and risk check-up; the same key order decides which key pays
DeepSeek and DeepInfra (upstreams serving our managed DeepSeek lane) The "DeepSeek" models above are served first-party by DeepSeek on a Company-held key, and DeepSeek is China-hosted; if a request to one of them fails, our gateway automatically retries it against DeepInfra, also on a Company-held key and hosted in the United States. DeepSeek V4 Pro has no failover and stays on DeepSeek. The same prompt/message content the DeepSeek lane processes — Ask Omniscio help questions, onboarding starter agents and Omniscio-credits coding sessions, only while no Google account is connected; never Google user data
CrofAI (retired coding lane) CrofAI ran as a Pro-gated coding lane on a Company key. It was retired on 2026-09-21 — its endpoint no longer serves requests, and no new session can select it — but it stays listed because existing CrofAI sessions and their recorded spend still resolve through it. CrofAI is China-hosted, so while such a session runs, the whole session — every prompt, every file the agent reads, and every tool result — goes there. The whole coding session's prompts, files and tool results, for a session that already uses the lane
GLM (Zhipu), Kimi (Moonshot), MiniMax, Meta Muse, Qwen (via our AI gateway, on Company-held keys) Cost-efficient coding lanes run on Company keys when you choose Omniscio credits: GLM goes direct to Zhipu's api.z.ai; Kimi / MiniMax / Meta Muse / Qwen go through the pooled Company OpenRouter key The prompt/message content those lanes process
FlowVoice (dictation — Omniscio cloud speech by default, or Google Gemini, Omniscio hosted, or Fluid Speech Mini on your own computer) OS-wide voice dictation (speech-to-text), on by default and engaged only while you hold its shortcut, and the screen recorder's automatic captions, which are on by default and are transcribed through Omniscio cloud speech whenever that is the engine (the default) — otherwise they transcribe on your own computer. Dictation uses the engine you choose in Settings: Omniscio cloud speech (the default) sends the audio through our cloud to Soniox, or to Inworld when Soniox refuses the stream or fails before any words are transcribed, then to Meta (Muse Voice Transcribe) if Inworld refuses too, and only where it is configured to Deepgram after that; Gemini transcribes in Google's cloud; Omniscio hosted sends it to our server, which streams it to Deepgram; Fluid Speech Mini transcribes on your own computer Live microphone audio you dictate, plus your dictation language; on Omniscio hosted only, the foreground app name too. For screen-recorder captions, the recording's audio — your microphone plus, by default, your computer's system audio, which on a recorded call or meeting includes other participants' voices — plus the recording language. With Gemini, that audio goes to Google — on your own Google key, or through our relay if you have none. The window title is not sent — it is captured locally for your own dictation history.
Supermail backend (Company self-hosted) Backend for the Supermail mail client — connecting a mailbox mirrors it server-side — and its contact enrichment, engaged only if you use Supermail. Contact enrichment sends a one-way (MD5) hash of a contact's email address to Gravatar to look up a public profile image Your mirrored mailbox: senders and recipients (incl. cc/bcc), subjects, full bodies, attachments, sent mail, drafts and labels; the mailbox's encrypted connection credentials; and contact-enrichment lookups
Omniscio backend apps (Company self-hosted) Our own backend services for support chat (the helpdesk), the plugin marketplace, Mission Control project boards (including the Monday.com and Trello board connections that run through them), importing boards from tools you connect, and the relay that carries some Shares — each engaged when you use that feature The support conversations, board items, members' names and email addresses, imported board data, marketplace requests and relayed Share content those features handle
unpkg.com Serves the React runtime into a viewer's browser for a Share whose document runs code — automatic whenever such a share is opened The viewer's IP address and browser user-agent; no user content
Hugging Face One-time download of the on-device embedding / captioning model weights the App runs locally Your IP address and browser user-agent; no user content
GitHub (ffmpeg-static release assets) Automatic, unprompted download of a pinned, checksum-verified ffmpeg static build when a recording first needs transcoding and the machine has no ffmpeg, after the package-manager route (winget / Homebrew) fails Your IP address and request metadata (to GitHub, and to the package sources behind winget / Homebrew); no recording or other user content
TypeSafe (api.typesafe.ai) The primary lane for the "Jev" decision feature — a typed question-and-answer call with a confidence score, run through our AI gateway on Company-held keys. It is the first leg of the chain, so it is reached by any signed-in user with AI credit rather than by an opt-in feature The situation text and the questions you submit to a decision call — free text that can contain names, email addresses, message bodies and agent output.
jsDelivr Serves files the Video Editing Suite plugin downloads to run on your computer — a Google Fonts typeface you pick, and the GPU runtime its on-device background removal needs — fetched only when the plugin needs one Your IP address and browser user-agent; no user content
ffmpeg build server (ffmpeg.martin-riedl.de) The Video Editing Suite plugin fetches a pinned ffmpeg static build from this server on macOS and Linux the first time it needs one and none is installed Your IP address and request metadata; no recording or other user content
Agent Client Protocol registry (cdn.agentclientprotocol.com) Download of the public list of coding agents you can add, when you open that list or add an agent (kept for a day) Your IP address and request metadata; no user content
Agent 37 (api.agent37.com) Omniscio's automated testing fleet rents disposable machines from Agent 37, on our account, and runs check jobs on them. It runs only when Omniscio's automated testing fleet is switched on for your work — no user action required, and nothing is sent while it is off. A machine that fails a check is stopped and its disk kept for inspection. The provider's processing region is not yet on file The repository checkout at the commit under test for an automated check run, that run's output and timings, and an opaque machine identity for the rented worker

Optional subprocessors (only when you enable a specific feature)

These run on Company-held keys, but only process data when you use the corresponding feature. Several are off by default.

Subprocessor Feature Data it may process
Klipy SMS GIF search in the SMS composer (uses a Company default key unless you set your own) Your typed GIF search terms
Firecrawl, Apify, X/Twitter Built-in web-scraping / data-extraction / read-only tweet-search tools The URLs, queries, or page content you request
OpenAI, Google Gemini, Perplexity ("bundled API keys," off by default) — not yet active Would let your agent call these providers through our gateway on Company keys. All three are pre-integrated but not funded, so no data reaches them today Once enabled: your prompts and context for those calls
DeepInfra (pooled-key coding lane) — not yet active A planned paid pooled-key lane that would run a whole coding session on a Company key. Staged, not funded — no data reaches this lane today; it will be moved to the Core table before being switched on. (DeepInfra is already live as our DeepSeek lane's failover — see its Core row.) Once enabled: the whole session — every prompt, file the agent reads, and tool result.
Fish Audio Default text-to-speech (read-aloud) on a Company-funded relay — engaged only if you turn on voice (off by default) The response text read aloud (may contain names or message content)
Soniox Text-to-speech (read-aloud / narration) on a Company-funded relay, alongside Fish Audio — engaged only if you turn on voice (off by default); and live speech-to-text on Omniscio cloud speech, the default engine — for default-on FlowVoice dictation while you hold its shortcut, and for the screen recorder's automatic captions, which are on by default The response or narration text read aloud, forwarded verbatim (may contain names or message content); for speech-to-text, the live microphone audio and your dictation language, and for screen-recorder captions the recording's audio — your microphone plus, by default, your computer's system audio, which on a recorded call or meeting includes other participants' voices
Inworld Live speech-to-text on Omniscio cloud speech — the fallback used only when Soniox refuses the stream or fails before any words are transcribed — for default-on FlowVoice dictation while you hold its shortcut and for the screen recorder's automatic captions, which are on by default For dictation, the live microphone audio you dictate and your dictation language; for screen-recorder captions, the recording's audio — your microphone plus, by default, your computer's system audio, which on a recorded call or meeting includes other participants' voices
Meta (Muse Voice Transcribe) Live speech-to-text on Omniscio cloud speech — the last resort among the shared engines, used only when Soniox, then Inworld, refuse the stream or fail before any words are transcribed — for default-on FlowVoice dictation while you hold its shortcut and for the screen recorder's automatic captions, which are on by default; also available as a speech-to-text engine on your own Meta / OpenRouter key, if you choose it For dictation, the live microphone audio you dictate and your dictation language; for screen-recorder captions, the recording's audio — your microphone plus, by default, your computer's system audio, which on a recorded call or meeting includes other participants' voices
Deepgram Live speech-to-text for FlowVoice dictation on Omniscio hosted, on a Company-held key — on by default, and engaged only while you hold its shortcut and have chosen that engine; and a last-resort engine on Omniscio cloud speech, used only where that engine is configured, where it also receives the screen recorder's automatic captions, which are on by default The live microphone audio you dictate and your dictation language; for screen-recorder captions, the recording's audio — your microphone plus, by default, your computer's system audio, which on a recorded call or meeting includes other participants' voices
Cerebras (through OpenRouter, on a Company-held key) The optional AI tidy-up of FlowVoice dictation on Omniscio cloud speech and Omniscio hosted — on by default, switchable in Settings; each request refuses data collection The text of each dictation (never the audio), from which filler words and stray repeats are removed before it is typed
JLS Image Studio (our own service) Image generation — an off-by-default, in-development feature. Sends your prompt and any reference images to our own image service, which forwards them to the image-model vendor it runs for the model you pick (Google Gemini, ByteDance Seedream, Ideogram, OpenAI gpt-image or Meta Muse image today, with Topaz, Bria and 851-labs for upscaling and background removal) Your prompt and your reference images (unredacted — they can contain people or other personal data), plus your account id, and the images, galleries and share links it stores on your behalf
Unclaimed-property check (our own service) The unclaimed-property lookup — off by default. Queries the public state unclaimed-property registries for the person you name The name, city and state you search for — a third party's personal data
Smrzz (our own service, smrzz.com, hosted on Google Cloud / Firebase) Summarizes what you send from the built-in Smrzz library (on for everyone from 6 November 2026; nothing is sent before you sign in): links, video captions, pasted transcripts and text, uploaded files, and the videos behind Omniscio Shares links you paste. It also receives your deletes and, if you use them, your followed YouTube channels. On the Omniscio-account path it is reached through our gateway on a Company-held key, and each summary is paid from your credits at the price shown in the app The links, captions, transcripts, text, files and shared videos you submit, your smrzz.com library, your deletes and followed channels, and (Omniscio-account path) your Omniscio account id
Jina AI (Reader, reached by smrzz.com, our own service) Reads web pages for the built-in Smrzz library: when you summarize a link that is not a YouTube, Vimeo, Loom, podcast-index or social-media video link, or a link to an audio or video file, smrzz.com asks Jina's Reader (r.jina.ai) to fetch that page and return its text. Spotify and Apple Podcasts links are read this way too The link you submit; Jina fetches that public page itself and returns its text. No account details are sent
SupaData (reached by smrzz.com, our own service) Reads social-media and media-file links for the built-in Smrzz library: for a Twitter/X, TikTok, Instagram or Facebook video link, smrzz.com asks SupaData (api.supadata.ai) for the post's title and author; for those links, and for a link straight to an audio or video file, it asks SupaData for a transcript only when smrzz.com cannot transcribe the media itself The link you submit; SupaData fetches that public post or file itself. No account details are sent
PodcastIndex (podcast directory, reached by smrzz.com, our own service) Looks up podcast episodes for the built-in Smrzz library: when you summarize a podcastindex.org episode link, smrzz.com asks PodcastIndex (api.podcastindex.org) for that episode's details and transcript location The podcast episode id of the episode you ask to summarize. The request comes from smrzz.com's server, not your device, so PodcastIndex does not see your IP address
fal.ai — not yet active An image-generation lane we have declared against a Company-held fal account. No part of the App sends anything to it today — it is wired but not called — and it will be moved into the Core list, with its terms confirmed, before it is ever switched on Once enabled: the image-generation prompt text you submit to that lane
ipwho.is (IP geolocation) Daily Weather Report — fills your place in automatically the first time, so you do not have to type a ZIP code, and the header Weather widget in the same way The public IP address of the machine running the App, from which the vendor derives a coarse location (country / region / city / postal code). No account, key, identifier or app content is sent. It runs once per app run, only while the weather report is switched on or the header widget is pinned, and never once you have typed a place yourself; the detected place is saved into your weather setting and you can edit it

Services you connect on your own account (not our subprocessors)

When you connect these on your own API key or account, data flows directly from your device to that provider under its terms — we are not the processor, so they are not our subprocessors. They are listed here for transparency:

How we notify you of changes

We maintain the current list on this page. When we add or replace a subprocessor, we will update this page and the effective date above, and — where required by an applicable data-processing agreement — give business/team customers advance notice and the opportunity to object as that agreement provides. Material changes are also reflected in the Privacy Policy's "Last updated" date.

Contact

Questions or a DPA request: legal@omniscio.com, Omniscio LLC, 5301 Terminal St, Charlotte, NC 28208.