Do AI meeting bots record video?
It varies. Audio is universal, because transcription needs it; video capture differs by tool and plan tier — some record the meeting screen including shared slides and demos, others skip video entirely. Donna records both: meeting audio and the screen, compressed after the call and stored on your own server.
Why does video capture matter for a meeting record?
Because meetings are not podcasts. Decisions get made about things on screen — a pricing table, a design, a dashboard, a contract clause — and a transcript that says “this one, not that one” is evidence of nothing. The screen recording preserves what was actually shown at the moment words were said, which is exactly what disputes and reviews need.
How does Donna capture and store video?
On the server, ffmpeg records the virtual display Donna’s browser runs in — the same view a participant sees — alongside the meeting audio from the virtual audio sink. After the call, the raw capture is compressed to a sensible size, muxed with the audio, and kept on your disk, with the meeting’s transcript and report referencing it in your dashboard. Nothing is uploaded anywhere.
Do participant camera feeds get recorded?
A bot records what a participant sees: the meeting layout, active speaker, and anything shared. There is no privileged backstage access — if a camera is off, there is nothing to record. This “sees what the room sees” property keeps bot recording ethically equivalent to a person recording their own screen, with visible attendance on top.
Related questions
Does video make recordings enormous?
Raw screen capture is large, which is why Donna compresses after every call — the stored artifact is a fraction of the raw size, tuned for “readable slides”, not cinema. Your disk, your retention policy.
Can I get audio-only operation?
Donna records the screen alongside audio in her live pipeline; for audio-only material, the upload path takes any existing audio recording and runs the same transcription and analysis.
Is the video included in the analysis?
The analysis runs on the transcript today; the video is kept as the visual evidence beside it. When a report timestamp points at 0:43:12, the recording shows what was on screen at 0:43:12 — which is usually the point.
Put Donna in your next meeting
Donna deploys onto your own VPS in an afternoon: nginx, pm2, PostgreSQL, your API keys. Early access is open — tell us about your team and we’ll get her a seat at your table.
Request early access