Capture a meeting
Understand microphone, shared audio and screen recording.
The meeting companion is a compact place to capture a conversation while keeping the rest of your desktop available. Capture options depend on the operating system and its WebView.
Understand the sources
Microphone is the input device selected for your own voice. Shared audio is the audio track supplied by the selected screen, window or tab. They are transcribed as separate channels when available.
A channel name is not a verified personal identity. If shared audio contains several people, Groq Whisper alone does not identify each individual speaker.
Headphones can reduce audio from speakers leaking into your microphone. Without them, a remote participant may also be audible on the microphone channel. Review source labels rather than assuming they establish who said something.
Start a recording
- Configure a transcription provider in Settings > AI & meetings.
- Open the meeting companion.
- Choose audio-only or Screen + audio as appropriate.
- Select the tab, window or screen. Enable audio sharing if the system offers it.
- Check the microphone and shared-audio indicators.
- Start recording, then test a short sample before a longer session.
Let meeting participants know you are recording. Avoid capturing unrelated windows or private material you do not intend to retain.
Screen recording and audio sharing
Selecting a screen determines the source. With screen recording enabled, Noter.ai also retains the visual recording for later playback. The operating system decides which shared-audio options are available; selecting a window does not necessarily include internal audio.
Linux live capture is unavailable in this candidate. macOS internal-audio capture is not guaranteed. See the Windows, Mac or Linux guide for your platform.
Transcripts and organized drafts
The transcript shows captured passages. Organized notes extract useful details, decisions, dates, requirements, changes and action items into a draft. Review important facts and ambiguous speech before saving.
If a provider falls behind, Noter.ai warns that some passages may be missing. The bounded transcription backlog should not keep Stop waiting indefinitely. The local recording remains the material to review for gaps.
Speaker recognition
Optional Deepgram or ElevenLabs connections can add speaker labels. They receive shared audio when normal usage permits that optional feature. Speaker numbers are local to the recognition result and can be imperfect; they are not verified names. With Groq alone, shared audio remains grouped.
End and review
Stop recording, open the saved meeting, review the transcript and organized draft, then save or update the linked note. Continue with recordings and transcripts.