Privacy
Where your words actually go.
Most privacy policies are written so that nothing in them can ever be wrong. This one is written so you can check it. Where a third party sees your data, it is named. Where something leaves your machine, it says so.
Last updated 24 August 2026.
The short version
- Your video and audio do not pass through our servers. Calls are peer-to-peer — the streams go browser to browser. Our server passes the messages that let two browsers find each other, and then gets out of the way. The exception is a recording, which is covered below.
- We do not train any model on your meetings, and we do not sell, rent or share your meeting content with anyone for advertising.
- There are no accounts, so we do not hold a password, a payment method, or a profile. The only thing you tell us about yourself is the name you type before you join.
- Transcription happens in your browser, not on our server — but in Chrome and Edge that involves Google. See below; this is the part people are most often surprised by, so it gets its own section.
What we hold, and for how long
| What | Where it lives | How long |
|---|---|---|
| Your display name | Server memory, for the meeting | Until the meeting is cleared |
| The transcript | Server memory | Until the meeting is cleared — about four hours after the last person leaves |
| Chat messages | Server memory | Same |
| Files you drop in chat | Our disk | Until deleted; anyone with the link can fetch them |
| Recordings | Our disk | Until deleted. Private recordings are refused to everyone without the owner key — the page and the video file both |
| Images generated in a meeting | Our disk | Until deleted |
| A QuickAlign record | Our disk, and listed at /records | Until deleted. This one is deliberately a public record of a decision |
| Developer keys | Our disk, as a SHA-256 hash | Until deleted. We cannot read your key back, only check one you present |
A meeting that nobody is in is dropped after about four hours, and with it the transcript and the chat. That is not a retention promise dressed up as a feature — it is how the room is built, and it means a meeting code is not a permanent open door.
Transcription, honestly
Each microphone is transcribed on the machine it belongs to, which is why the speaker on every line is known rather than guessed. How that happens depends on your browser, and the difference matters:
- Chrome and Edge use the browser's built-in Web Speech API. In Chrome, that API sends audio to Google's speech servers for recognition. It is Google's processing, under Google's terms, and neither we nor you can see inside it. We use it because it is fast, free and supports two dozen languages — but you should know it is happening.
- Safari and Firefox run a Whisper model inside the tab. Nothing leaves your machine. It is currently English-only, and the setting says so rather than failing quietly in another language.
- Either way, the resulting text goes to our server, because the room is built out of it — the live notes, the agenda, the assistant and the search all read the same transcript.
- You can turn it off. Settings → Transcription. Your voice is then not written down, and the notes will not include what you say.
Who else sees it
Named, because "trusted third-party service providers" tells you nothing.
| Anthropic | Transcript text, to write the live notes, the recap, the agenda and the assistant's answers. Anthropic's API terms do not permit training on API inputs. |
| Audio, if you are transcribing in Chrome or Edge — via the browser's own speech API, as above. | |
| Fish Audio | The text of what the assistant is about to say, to turn it into speech. Also audio, if you ask us to transcribe an existing recording. |
| fal.ai | An image prompt, when a picture is generated. The prompt is written from the conversation. |
| Fly.io | Hosts the server and the disk. Our data sits on their infrastructure. |
| Google, Twilio | STUN servers, used to help two browsers find each other. They see network addresses, not content. |
That list is exhaustive as of the date at the top. If it changes, this page changes.
Agents
A program can join a meeting with a key. Everything about how we did that is a response to how badly it goes elsewhere:
- An agent is in the member list, like a person. There is no hidden listener.
- Its arrival is written into the chat, where everyone sees it.
- Anyone in the room can remove it — not just the host.
- The host can turn agents off entirely with one switch, and every connected agent is disconnected and told why.
- An agent receives the transcript and the chat. It does not receive raw audio or video.
Recording
A recording is made in the browser of whoever pressed record — their machine composites the tiles and mixes the audio — and uploaded to us in pieces while the meeting runs. Everyone in the room sees a recording indicator for as long as it is running. There is no silent recording.
Visibility is set by the person who made it: private (only someone holding the owner key can open it, and the video file itself is refused to everyone else), anyone with the link, or public and listed. Laws about recording people vary by where everyone is sitting, and complying with them is the responsibility of the person recording.
What we do not do
- We do not train models on your meetings.
- We do not sell or rent your data.
- We do not run advertising, and nobody pays us to put anything in front of you.
- We do not use tracking cookies or third-party analytics. The only thing we keep in your browser is your own settings — your name, your language, your camera choice — and that never leaves your device.
- We do not read your meetings. The transcript is processed by software to build the notes; no person here is reviewing it.
Your choices
Because there are no accounts, most of the usual rights resolve to something simple: leave the meeting and the room forgets you within hours. For anything that persists — a recording, a file, a QuickAlign record — write to the address below with the link and we will delete it. If you hold the owner key for a recording you can delete it yourself.
If you are in the EU or UK, our legal basis for processing meeting content is legitimate interest in providing a meeting you asked to be in. If you would rather we did not process what you say, turn transcription off before you speak.
Children
Team Meet is not directed at children under 13, and we do not knowingly collect their information. A school using this should be the one deciding whether that is appropriate, with consent from the people it needs consent from.
Contact
Questions, deletions, or a claim that something on this page is wrong: privacy@teammeet.io. The last one is the most useful kind of mail we can get.