Stay in the meeting; the notes write themselves
You can't fully participate and take good notes at the same time. Nod captures the meeting from your Mac's audio and builds the recap for you, so you keep your attention on the conversation and still leave with the follow-ups straight.
- Works across Zoom, Google Meet, Teams, and calls on speaker
- Structured recap, not a raw transcript dump
- Action-item text can include a named person or deadline; review it against the call
- Every meeting becomes searchable with 'Ask Nod'
No bot, no stored audio
Nod never joins the call as a participant, and it keeps no recording. Capture is local on your Mac; transcription isn't. Each audio segment is sent through a European proxy to hosted transcription, and Nod discards it when the text returns. No model trains on the data. Groq may retain transient error or abuse logs for up to 30 days unless Zero Data Retention is enabled. Only the transcript and summary are saved by Nod, encrypted at rest in the EU (AWS eu-west-1). See the security page.
How bot-free capture works, and how it compares with a bot in the call, is covered in meeting notes without a bot: the complete guide. For a per-platform walkthrough, see AI notes in Google Meet without a bot.
From agenda to minutes
Pair structured notes with a clear plan: see how to write a meeting agenda going in and how to write meeting minutes coming out.
Frequently asked questions
How does Nod take AI meeting notes?
Nod captures the meeting from your Mac's system audio, transcribes it, and structures it into topics, decisions, action-item text that can include a named person or deadline, and open questions — automatically, with no bot in the call. Review generated attribution against the source conversation.
Does a bot join the meeting?
No. Nod captures your Mac's audio rather than dialing in, so nothing appears in the participant list. You should still get consent where recording law requires it.
Are my meetings recorded and stored?
No audio is stored. It's held in memory about five seconds to transcribe, then discarded — only the transcript and summary are saved, encrypted in the EU with no model training.