A free AI note taker that runs in your browser
OpenNoteTaker (opennotetaker.app) is a free AI note taker for meetings. It records the call from your browser tab, so no bot joins, and it transcribes in the page, so the audio is never uploaded. The work happens on your own hardware, which is why it costs nothing: not a trial, not a free tier with a timer, and no account needed.
What this free AI note taker does, at no cost
| Feature | Where it runs | Price |
|---|---|---|
| Recording, with no bot in the call | Your browser | Free |
| Transcription, 99 languages | Your browser | Free |
| A timestamp on every sentence | Your browser | Free |
| Telling speakers apart | Your browser | Free |
| Topics, key lines, action items, decisions | Your browser | Free |
| Search across every meeting | Your browser | Free |
| SRT, WebVTT, text, Markdown, JSON, CSV, HTML | Your browser | Free |
| Noticing the Meet, Zoom, Teams or VooV call you joined | An optional extension | Free |
The summary is extractive: it pulls the topics, key lines, action items and decisions out of the transcript itself, in your browser, rather than asking a language model to write prose.
What is not free, and why
Three features run on our backend with our model key, so they use credits: AI-written minutes, a question across your whole library, and translating a finished transcript into any language. They send transcript text, never audio. The price of each job is shown on the button before you press it, credits do not expire or renew, and a job that fails costs nothing. If you never press those buttons, the free AI note taker is all you use. See pricing.
Why a free AI note taker can stay free
Transcription is Whisper compiled to WebAssembly, running in the tab; speaker separation, the summary and the exports are Rust compiled to WebAssembly. The model is downloaded once from Hugging Face and cached, and after that transcription works offline. Your notes live in your own browser storage, and the audio is discarded when you close a note unless you choose to keep it. There is no server of ours doing the work, so there is nothing to charge a subscription for.
Before you rely on it
- Chrome or Edge for calls. Tab audio — which is everyone else on the call — can only be captured in Chrome and Edge, and only when you pick the tab and tick “share audio” in the browser’s share dialog.
- Desktop meeting apps. Zoom and Teams installed as applications are not browser tabs. Join from the web client, or route the sound through a loopback device such as BlackHole or VB-Cable.
- Speakers. Separation works well for two or three distinct voices and struggles with similar voices and crosstalk. You can correct any line.
- Consent. Every recording opens on a consent screen that cannot be skipped. The free AI note taker cannot tell you whether you may record; the people in the call can.
OpenNoteTaker is an open-source AI note taker under MIT or Apache-2.0, with the source on GitLab. For a specific platform, see the guides for Google Meet or for Microsoft Teams.
Questions
Is this free AI note taker really free?
Yes, for everything that runs on your machine: recording, transcription, speakers, the summary, search and every export. Only AI-written minutes, library questions and translating a finished transcript use credits.
Do I need an account?
No. An account only holds a sign-in method and a credit balance, and you only need one for the three credit features.
Does it join my meeting as a bot?
No. It records the audio playing in your browser tab, so nothing appears in the participant list.
Does it work offline?
After the first visit, yes: the transcription model is cached, and transcription works with the network off.