Legal
Privacy Policy
This is the plain-English version of how Coommit handles your data. We've kept it short on jargon and long on specifics. The same document, in legal form, is what governs your rights.
Effective May 15, 2026 · Last updated June 11, 2026
1. The basics
Coommit is a collaborative meeting workspace operated by TAAO, Inc., a Delaware C-Corporation. When we say "we," "us," or "Coommit" in this document, we mean TAAO, Inc.
You can reach us anytime at hello@coommit.com or by mail at 1007 N Orange St, 4th Floor 5510, Wilmington, DE 19801, USA.
This Privacy Policy covers our marketing site at coommit.com and our product at app.coommit.com (together, the "Services").
2. Our core privacy commitments
Coommit is built around seven commitments. The rest of this policy explains how we keep them.
- We never sell your data. Not to advertisers, brokers, or anyone else. Ever.
- We don't train our own AI models on your content. Coommit doesn't operate proprietary models. We orchestrate calls to providers you choose.
- You bring your own keys (BYOK). Your AI prompts go directly from our servers to your chosen provider, signed with your own API key. We hold the key encrypted; we never use it for anything besides routing your request.
- We delete inactive accounts after 12 months. If you stop using Coommit for a year, we erase your content.
- You can export and delete everything at any time. One click to delete your account, instant hard-delete cascade. No 30-day "grace" purgatory.
- No advertising, no advertising cookies, no third-party trackers beyond what's strictly needed. We use Google Analytics for product metrics. That's it. No ad networks.
- We're a small team and we know what we do with your data. If something below is unclear, email us and a real person will answer.
3. What we collect and why
3.1 Account information
When you sign up: your email address, name (if provided), a hashed password, your locale, your plan, and your subscription status. You may optionally connect Google or GitHub OAuth, in which case we receive your name, email, and avatar URL from those providers.
3.2 Content you create
Anything you put into Coommit: canvas elements (text boxes, images, embeds, sketches, tasks), chat messages, meeting transcripts, Loom-style recordings, Echo brain notes, files you upload, and your room settings. This is your content; you own it.
3.3 Calls and meetings
Audio: live call audio is streamed through our servers to Google Cloud Speech-to-Text for real-time transcription. The raw audio is never persisted on our infrastructure. It transits in memory and is discarded.
Transcriptions: the resulting text is stored in our database, attached to the room, until you delete it (or the 12-month retention runs).
Recordings: if you create a Loom-style recording, the video file is stored in our private cloud storage (Belgium region) and only accessible via short-lived signed URLs.
3.4 BYOK API keys
If you connect an Anthropic, OpenAI, or Google Gemini API key for Echo (our AI teammate), we store it AES-encrypted in our database. We only decrypt it in memory, briefly, to forward your request to the provider. We never log it, never share it, and never use it for anything you didn't initiate.
3.5 Integration tokens
If you connect Google Calendar or Calendly, we store the OAuth tokens needed to read and create events on your behalf. You can disconnect at any time from Settings. We hard-delete the tokens immediately. For everything specific to Google user data — including our Limited Use disclosure — see Section 14.
3.6 Usage and analytics
We collect aggregated usage metrics: page views, feature usage, error reports, performance data. We use this to fix bugs and improve the product. See §6 for the tools involved.
3.7 Acquisition data
When you visit our marketing site or start a free trial, we log: referrer URL, UTM parameters, Google Click ID (gclid), approximate country (derived from IP; we don't store the IP itself long-term), and the user agent. This helps us understand how people find Coommit. Stored in your account record.
3.8 Communications
Emails you send us at hello@coommit.com, support conversations, and your responses to product surveys.
4. AI providers and BYOK
Echo, Coommit's AI teammate, is powered by third-party large language models. Coommit doesn't train any AI models. We orchestrate calls to providers you choose, using your API keys.
4.1 How BYOK works
- You paste your API key into Settings → Echo → BYOK form.
- We encrypt it with AES (server-side secret) and store it in our database.
- When Echo runs, we decrypt your key in memory, send your request to the provider, and discard the key from memory.
- Your prompts (transcripts, canvas state, query) and the provider's response transit through Coommit's servers for orchestration only. We don't store the prompts beyond the brain-note summaries you opt into.
- If you haven't configured a key, Echo refuses to run. There is no fallback to a Coommit-owned key.
4.2 Supported providers
- Anthropic (Claude models): Anthropic's commercial terms commit to not training on customer inputs or outputs.
- OpenAI (GPT models): OpenAI's API data-usage policy commits to not training on data submitted via API.
- Google Gemini: Google's paid API tier commits to no training. Important: if you use a free-tier Google AI Studio key, Google may use your prompts to improve their models, per Google's own terms. We show a warning when you paste a Gemini key, but the decision is yours.
4.3 What we send to providers
When Echo is invoked: the relevant transcript fragments, the current canvas state, your query, and our system prompt. Sent over TLS to the provider's API. Coommit doesn't add identifiers. The provider only sees what your key authenticates.
4.4 Other AI-related processing
Speech-to-text (Google Cloud STT) and text-to-speech (Google Cloud TTS) are run with our enterprise contracts, which prohibit training on customer audio.
6. Sub-processors we rely on
We rely on the following third parties to operate Coommit. They process your data on our behalf, under contracts that bind them to confidentiality and security standards comparable to ours.
6.1 Hosting and infrastructure
| Vendor | Purpose | Region |
|---|---|---|
| Google Cloud Platform | Application hosting (Cloud Run), database (Cloud SQL PostgreSQL), file storage (Cloud Storage), logging, monitoring, build pipelines, BigQuery (billing analytics) | us-central1 (app), europe-west1 (database & file storage), europe-west9 (Gemini Live) |
6.2 Real-time meetings & voice
| Vendor | Purpose |
|---|---|
| Metered | TURN / STUN servers for WebRTC NAT traversal during calls |
| Google Cloud Speech-to-Text | Live transcription of meeting audio |
| Google Cloud Text-to-Speech | Echo's spoken responses (voice synthesis) |
6.3 AI providers (BYOK: you choose)
| Vendor | Purpose |
|---|---|
| Anthropic | Claude models (Echo chat); your key, your account |
| OpenAI | GPT models (Echo chat); your key, your account |
| Google Gemini | Gemini models & embeddings (Echo chat & memory); your key, your account |
6.4 Authentication
| Vendor | Purpose |
|---|---|
| Google OAuth | "Sign in with Google" login |
| GitHub OAuth | "Sign in with GitHub" login |
6.5 Payments
| Vendor | Purpose |
|---|---|
| Stripe | Subscription billing & payment processing (Stripe receives your name, email, and payment information; Coommit never sees your card details) |
6.6 Email, monitoring & integrations
| Vendor | Purpose |
|---|---|
| Resend | Transactional email (account verification, invites, password reset, notifications) |
| Sentry | Error and performance monitoring |
| Google Analytics 4 | Aggregated usage analytics on the marketing site and product |
| Browserbase | Powers the optional Canvas Browser Box (a shared headless Chrome session inside a room) |
| Calendly | Optional integration if you connect your Calendly account (OAuth + webhooks) |
| Google Calendar API | Optional integration if you connect your Google Calendar (OAuth) |
| Google Drive Picker API | Optional file picker if you import from your Google Drive |
| country.is | Fallback IP-to-country resolution for acquisition analytics (no IP is stored) |
6.7 Marketing site (coommit.com) only
| Vendor | Purpose |
|---|---|
| Spline | 3D scene rendering on the homepage hero (CDN-hosted) |
| Masko | Echo mascot animated video (CDN-hosted) |
| FormSubmit | Relays the contact form on the Contact page to hello@coommit.com |
An always-current list is also maintained internally and can be requested at hello@coommit.com.
7. Where your data lives
The Coommit app currently runs from Google Cloud's us-central1 region (Iowa, USA). Your database records (account, rooms, transcripts, brain notes, recordings metadata) and your file storage (canvas assets, recordings, avatars) live in Google Cloud's europe-west1 region (Belgium).
This means: parts of your data live in the EU by default, even though our application servers are in the US. When you initiate an Echo request to a US-based AI provider (Anthropic, OpenAI), your prompt travels from our servers to theirs over TLS, a transatlantic transfer governed by Section 13 below.
8. How long we keep your data
8.1 Inactive accounts
If you stop using Coommit for 12 consecutive months (no logins, no activity, no API calls), we will email the address on file and, 30 days later, hard-delete your account and all associated content.
8.2 Active accounts
While you use Coommit, we keep your content for as long as you keep it: rooms, transcripts, recordings, brain notes, files. You can delete any of these individually at any time.
8.3 When you delete your account
Account deletion (Settings → Delete account) triggers an immediate hard-delete cascade: rooms you own, transcripts, chat messages, recordings, brain notes, file uploads, friendship and notification records, BYOK keys, integration tokens, all removed. There is no soft-delete grace period.
8.4 Specific retention windows
| Data type | Retention |
|---|---|
| Raw call audio | Not stored. Streamed through memory only. |
| Auth session cookies | 30 days (rolling) |
| Waitlist session cookie | 90 days |
| Guest invite link | 7 days |
| Guest session cookie | 24 hours |
| Signed URLs for files | 1 hour (recordings), 6 hours (canvas assets), 7 days (avatars, bug-report screenshots) |
| Echo transcript-summary cache (in-memory) | 30 minutes |
| Brain digest cache (in-memory) | 5 minutes |
| Billing & tax records (legal obligation) | Up to 7 years after the last transaction, as required by US and EU tax law |
| Backups | Encrypted snapshots retained up to 35 days; deletions propagate within this window |
10. Your rights
10.1 If you're in the EU, UK, EEA, or Switzerland (GDPR / UK GDPR / FADP)
You have the right to:
- Access: get a copy of the personal data we hold about you.
- Rectify: correct anything inaccurate.
- Erase: ask us to delete your data ("right to be forgotten"). You can also do this yourself from Settings.
- Restrict: limit how we process your data.
- Object: to processing based on legitimate interests, including analytics.
- Port: receive your data in a structured, machine-readable format.
- Withdraw consent: for processing that relies on consent (e.g., optional integrations).
- Lodge a complaint: with your local data-protection authority.
Our legal bases under GDPR: performance of a contract (operating Coommit for you), legitimate interests (security, fraud prevention, product analytics), consent (optional integrations, marketing emails), and legal obligation (tax records).
10.2 If you're in California (CCPA / CPRA)
You have the right to:
- Know what personal information we collect, use, share, and sell (we don't sell).
- Delete the personal information we hold about you.
- Correct inaccurate personal information.
- Opt out of "sale" or "sharing": moot here, because we do neither.
- Limit use of sensitive personal information.
- Non-discrimination: we won't penalize you for exercising any right.
10.3 If you're in another US state with a privacy law
(Virginia, Colorado, Connecticut, Utah, and others.) You have analogous rights under your state's law. The fastest way to exercise any of them is the same as above: email us.
10.4 How to exercise your rights
Email hello@coommit.com from the address tied to your account, or use the Settings page in the product. We respond within 30 days (sometimes faster). We may ask you to verify your identity if the request is sensitive.
You can also delete your account yourself, instantly, from Settings → Delete account.
11. Children's privacy
Coommit is intended for users aged 13 and over.
For users in the European Economic Area, the UK, or Switzerland, the digital age of consent is typically 16 (some member states have lowered it to as young as 13). If you are below the digital age of consent in your country, you may only use Coommit with the consent of your parent or legal guardian.
We do not knowingly collect personal data from children under 13. If you believe a child under 13 has provided data to Coommit, email hello@coommit.com and we will delete it.
12. Security
We take security seriously. Concretely:
- Encryption in transit: TLS 1.2+ for everything between your device and our servers, and between our servers and our sub-processors.
- Encryption at rest: Google Cloud's default storage encryption for the database and file storage.
- BYOK keys: additionally encrypted at the application layer with AES, using a server-side secret that is itself rotated.
- Hashed passwords: passwords are never stored in plaintext.
- Hard deletes: when you delete content, it's gone (subject to the backup window in §8.4).
- Access controls: only a small team of TAAO, Inc. employees has production access, on a need-to-know basis.
- Monitoring: Sentry alerts us to errors and anomalies in real time.
- Coordinated disclosure. Found a vulnerability? Email hello@coommit.com. We respond fast, we credit researchers who want it, and we don't sue.
No system is perfect. If we ever experience a security incident affecting your data, we will notify affected users without undue delay and, where required, the relevant data-protection authority within 72 hours.
13. International data transfers
Coommit is a US company with infrastructure in both the US (us-central1) and the EU (europe-west1, europe-west9). When data moves from the EU to the US (for example, when you use an AI provider hosted in the US), we rely on:
- Standard Contractual Clauses (SCCs) approved by the European Commission, with our sub-processors.
- EU–US Data Privacy Framework certifications, where our sub-processors are certified.
- Your explicit consent. When you connect a BYOK key to a US provider, you authorize the transfer in the act of doing so.
14. Google user data & API Services (Limited Use)
If you connect Google Calendar or import files from Google Drive, Coommit accesses Google user data through Google APIs. This section is our affirmative statement of how that data is handled.
14.1 What we access
- Google Calendar (
calendar.events): we read the events on your calendar so you can choose which ones to link to a Coommit room, and we create or update events on your behalf — for example to add a room link to a meeting or to put a call scheduled in Coommit on your calendar. - Google Drive (
drive.file): we only access the specific files you explicitly select in the Google Drive picker, to embed them into your collaborative canvas. This scope does not let us see the rest of your Drive. - Google account email (
userinfo.email): to display which Google account is connected.
14.2 How we use it — and how we never will
Google user data is used solely to provide the user-facing features described above, at your request. We do not use Google user data for advertising of any kind (targeted, personalized, retargeted, or interest-based), do not sell it to data brokers or anyone else, do not use it to determine creditworthiness or for lending purposes, and do not transfer it to third parties except as necessary to provide these features (the hosting sub-processors in Section 6), to comply with applicable law, or as part of a business transfer as described in Section 9.
14.3 Limited Use disclosure
Coommit's use and transfer to any other app of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.
14.4 No AI/ML training on Google user data
We do not use any data obtained through Google Workspace APIs — whether raw, aggregated, anonymized, or derived — to develop, improve, or train generalized or non-personalized artificial-intelligence or machine-learning models. Coommit trains no models of its own (Section 4), and Google user data is never used as training material for anyone.
14.5 Storage, retention and deletion
We store the OAuth tokens needed to act on your behalf, plus the minimal linkage records that map a calendar event to the Coommit room you created for it. They live in our database, encrypted at rest by our cloud provider. Disconnecting the integration from the Integrations page revokes the token with Google and immediately hard-deletes it from our database; deleting your account removes everything else (Section 8).
15. Changes to this policy
If we materially change this policy, we'll update the "Last updated" date at the top and email registered users at least 30 days before the changes take effect. Minor edits (typo fixes, link updates) won't trigger a notice but will be reflected in the date.
Past versions are available on request.
16. Contact us
Privacy questions, data requests, or anything else: write to a real human.
Or by mail to our registered office:
TAAO, Inc.1007 N Orange St, 4th Floor 5510
Wilmington, DE 19801
USA