Privacy at Trippi Notes

1. What the extension listens to

Only the sound of the meeting tab, and only after you click the extension icon in that tab. Chrome gives tab audio through the activeTab permission, which needs that click every time. The extension cannot start listening on its own, and it never touches your microphone or any other tab.

While it is listening, the panel shows it. Closing the panel or pressing stop ends the capture.

2. Where the audio goes

The audio is streamed from your browser to our speech-recognition provider over an encrypted connection, to be turned into text. It is not recorded or stored, by the provider on our behalf or by us. The connection uses a short-lived session key that our server issues for that one session.

If you turn translation on, each finished line is sent through our server to a machine-translation provider and comes back translated. It is not written to our database or our logs.

If you add terms to the glossary, they are sent with the audio stream so names and product words are spelled your way.

3. Where the transcript lives

The transcript, its translation, the speaker names, your own notes, summaries and tasks are kept in your browser’s local storage, on your computer. When you are signed in to your Trippi account, a copy of each meeting is also kept in the meeting archive on our server (see section 4), which is what makes the meetings journal in your account work on any computer. That journal is yours: you can delete one meeting or all of them from the journal.

4. When the text of a meeting reaches us

Only in these cases, and each one starts with something you do:

• Summarize, Ask AI and the follow-up email. When you press one of these, the text of that meeting (and your glossary) is sent to our server, which asks a language model and returns the result. The text is used for that one request and then dropped: it is not written to our database and not written to our logs.
• The email after a call. If you are signed in and leave this option on, the transcript of a meeting is sent to your own account’s address when the call ends. It can be switched off in the settings.
• The meeting archive. When you are signed in, it is on by default: when a meeting ends, it is stored on our server so you can open it, search it and ask questions about all your meetings from any computer. You can turn it off in the settings or in your account, and from then on no new meetings are stored. Questions about all your meetings are answered the same way as Ask AI: the relevant parts of your archived meetings are sent to a language model for that one request and are not kept by it. You choose how long they are kept (30, 90, 180 or 365 days; 90 unless you change it), and you can delete one meeting or all of them at any time.
• A link to a meeting, if you make one. This needs the archive. Anyone with the link can read that meeting until the link expires (30 days unless you choose otherwise, at most a year) or until you turn it off. On a team plan you can limit the link to your team: then only a signed-in member of your team who can already see the meeting opens it.
• Your team, if you are on a team plan. People on the same Trippi Notes team can read a meeting of yours from the archive only when you open it to them: with the setting that makes all meetings visible to the team, by giving one colleague access, or by adding it to a shared team space. They then see its summary, tasks and transcript, can export it, and see the comments left on it. A meeting you mark private is closed to everyone on the team, whatever the team settings say; only an open link you made yourself still works until you turn it off. Prompts for the AI chats that someone shares with the team are visible to the whole team. The owner of the team also sees each member’s email and how many meetings and hours each person had in the last 30 days, not what was said in them. People outside your team never see any of this.

Buttons that open ChatGPT or Claude hand the meeting to that service in your own browser; what happens there is governed by that service.

5. What else we keep

• An install record. When the extension is installed it registers with our server and gets an anonymous identifier and a token. No name and no email until you sign in.
• Counters. How many minutes were recognised and how many summaries and answers were used, per month. This is what your plan is measured against.
• An account, if you make one. Your email address, linked to your installs, so a plan works on any computer you sign in on.
• Usage events. Whether the transcript started, on which meeting platform, why it failed and how long a session ran. Never the text, meeting names or the pages you visit. You can turn this off in the settings.
• The address a request came from. As any server does. We keep it for at most 60 days and only to stop the free allowance from being reset by reinstalling the extension.
• Task trackers you connect. If you connect Linear, Jira or Asana, we keep the access token that service gave us (encrypted), the name of the workspace or site, and where your tasks go. Disconnecting deletes all of it.
• Keys for AI assistants. If you create a key to use your journal from Claude, ChatGPT or another assistant, we keep only a fingerprint of the key and when it was last used. Anyone holding the key can read your journal and manage your tasks until you revoke it in the journal.

6. Payments

Plans are sold through Paddle, the merchant of record. Card details are entered on Paddle’s side and never reach us; we receive the fact of a subscription and its state.

7. Who else sees anything

• The speech-recognition provider (Soniox): the audio stream, to turn it into text.
• The machine-translation provider: finished lines, if translation is on.
• The language-model provider (Anthropic Claude, Groq or Google Gemini): the meeting text, when you ask for a summary, an answer or an email.
• Our email provider: the address and the message, for the email after a call.
• Paddle: billing data, if you buy a plan.
• People on your team, if you are on a team plan: only the meetings you open to them, and the comments on those meetings (section 4).
• The task tracker you connect (Linear, Jira or Asana): each task you send there, with its owner, due date and a link to the meeting. Only when you send it, or when you turned on sending accepted tasks automatically.
• The AI assistant you give a key to: the meetings and tasks it asks for with that key.
• Yandex Metrica, on this website only: this site (not the extension) counts visits with Yandex Metrica, the same counter as the rest of gettrippi.app. It sees which pages of the site were opened and which links were clicked. Nothing from the extension and no text of your meetings ever reaches it.

They process what they receive to deliver the service and act as processors for us. Nobody else. We do not sell data and we do not run advertising.

8. What we never do

• Record or store the audio or video of your meetings.
• Join your call as a participant or a bot.
• Read tabs other than the meeting tab you started it in.
• Store your meetings on our server when you are not signed in or have turned the archive off.
• Show a meeting you marked private to your team.
• Use your meetings to train models.

9. Your rights

Write to [email protected] to get a copy of what is stored about your account or to have it deleted. Deleting the account removes the account record, the installs linked to it, their counters and any archived meetings.

Removing the extension stops all processing at once, and removes the journal kept in your browser.

10. Changes

If the product starts doing something this page does not describe, this page changes in the same release, not after it. The date at the top is the date of the last change.