AI meeting notes
Turn a meeting transcript into notes you can act on.
Review a timestamped transcript, then request a focused summary, decisions, and next steps.
In short: Meeting Recorder creates a timestamped transcript in your browser, then can turn that transcript into a summary, decisions, and action items. The audio stays local in the default Whisper workflow; only transcript text is sent to the configured summary provider when you request AI notes. Review the source before exporting Markdown, TXT, JSON, or PDF.
Open the browser recorder Review the transcription workflow
How to use it
- 01 — Record or upload
Use the browser recorder or upload a meeting file you are allowed to process.
- 02 — Transcribe locally
Run Whisper in the browser, then review timestamps and adjust speaker labels where needed.
- 03 — Request and review notes
Ask for a summary, decisions, and action items only after checking that the transcript is usable.
Example meeting notes
Product planning review
Summary: The team chose onboarding improvements as the next sprint focus and moved the research review to next month.
Decisions: Prioritize onboarding improvements · Review research next month
Action items: Priya — update help center content · Alex — share the latest onboarding flow
Frequently asked questions
Does AI meeting notes upload my audio?
Not in the default local workflow. Whisper runs in the browser and the audio stays on this device. When you request AI notes, only the transcript text is sent to the configured summary provider.
Can the notes identify speakers automatically?
The transcript includes speaker labels that you can review and edit. Local Whisper does not promise reliable automatic speaker diarization, so verify names and labels before sharing.
What do the AI notes contain?
The current workflow returns a concise summary, decisions, and action items. Owners are only used when stated in the transcript; otherwise the result uses Unassigned.
Can I correct the notes before exporting?
Yes. Review the transcript, timestamps, speaker labels, decisions, and action items before exporting Markdown, TXT, JSON, or print-ready PDF.
What if the AI summary is not configured?
You can still review and export the local transcript. AI notes depend on the configured summary provider and may be unavailable when no provider key is configured.
Start with a transcript
Record with the browser microphone, capture a permitted tab or window, or upload MP3, WAV, M4A, or MP4. Local Whisper produces the source transcript before notes are requested.
Separate decisions from follow-ups
The notes workflow keeps the summary, decisions, and action items distinct. Owners are retained when the transcript names them; otherwise the result uses Unassigned rather than inventing a person.
Use timestamps as evidence
Timestamped transcript lines let you return to the relevant audio and correct a note before sharing it. Speaker labels can be reviewed and edited in the workspace.
Know where AI processing happens
Local transcription does not upload the audio. When you request AI notes, the transcript text is sent to the configured summary provider; the page does not silently send or auto-join a meeting.