Speakers and voices

MeetCap tells the people in a meeting apart by their voices and shows each one as a speaker, such as Speaker 1 and Speaker 2. When you give a speaker a name, MeetCap remembers that person’s voice and recognizes them in later meetings.

To name someone, see Name and recognize speakers. This page explains how naming and recognition work.

How MeetCap tells speakers apart

MeetCap records your microphone and all audio on this computer separately, so it doesn’t mix up the people in the room with the people on the call.

During the meeting, the captions show a speaker for each line as it comes. When you stop, MeetCap identifies the speakers again from the whole recording, so the speaker of some lines may change. A line it can’t match to a speaker yet shows Unknown speaker.

MeetCap tells apart up to 8 speakers through the microphone and up to 8 in all audio on this computer. If more people speak, two of them can show as one speaker.

Naming a speaker

A name you give applies to this meeting and to the meetings that come after it. If one person shows up as two speakers, give both the same name, and they become one person in this meeting.

After the meeting, the AI that writes your minutes can suggest who a speaker is, from what was said. The suggestion looks like “Speaker 2 may be Alex (based on 3 lines)”. A suggestion isn’t applied until you confirm it. To apply it, click Confirm.

To change a name in one meeting, rename the speaker in that meeting. MeetCap changes the name there and remembers the voice under the new name. Names in other meetings stay as they are.

Remembering a voice

When you name a speaker, MeetCap saves a few clips of their voice in your meeting library. It needs about 30 seconds of that person speaking clearly in the meeting. If they spoke less, the name still changes, but the voice isn’t remembered. Name them again in a later meeting.

MeetCap needs the meeting’s recording to remember a voice. If you name someone on a Mac that didn’t record the meeting, or after the recording was cleared, only the name changes.

Recognizing people in later meetings

When a meeting ends, MeetCap compares each speaker with the voices it knows. If one matches, it uses that name. A name you give yourself always comes first. Someone who says only a few words in a meeting isn’t recognized.

Because voices are kept in the library, every Mac that uses the same library recognizes the same people.

If MeetCap already knows someone whose voice is very similar, it tells you when you name a speaker. If they are the same person, rename one person to the other’s name in Settings › People, and MeetCap merges them. Merging can’t be undone.

Where voices are kept, and how to delete one

MeetCap keeps voices in the hidden .meetcap folder in your meeting library. It never sends them to an AI service, and it doesn’t save voices of speakers you haven’t named.

To delete someone’s voice, open Settings › People, click Delete next to their name, then click Delete again to confirm. MeetCap deletes the voice and its clips right away and stops recognizing that person. Their name stays in past meetings.

To rename someone in every meeting at once, click Rename in Settings › People.

Tell people before you save their voice

A saved voice can identify a person, and many places treat it as sensitive personal information. Before you name someone, tell them that MeetCap will save their voice, where it is kept, and how to delete it. Get their consent if the law where you are requires it. See the privacy notice, section 5.