Why does my transcript say Speaker 1 and Speaker 2?

The app separated the voices but has no way to know whose they are. It either couldn’t read names from the meeting app, or it doesn’t remember voices from earlier meetings.

How does an app know who is speaking?

One of two ways. It copies names from the meeting app, the call window or the calendar invite, or it recognizes how each voice sounds. The first needs a supported call app with everyone joined under their own name. The second works in any app or in a room.

Do I have to name speakers in every meeting?

Not with an app that learns voices. In Enzo, click a speaker’s name once to say who it was, and Enzo names that person in later meetings.

Can speaker names work for people in the same room?

Yes, if the app tells people apart by voice. A meeting app sees one shared microphone as one participant, so names taken from the meeting app can’t separate them.

Where are the voices kept?

It depends on the app; some keep voice data in their cloud. With Enzo, who is speaking is worked out on your Mac, and recordings, transcripts and the voices it has learned stay there.

Guide

How to Get a Meeting Transcript With Speaker Names

Name: Enzo
Author: Enzo

Why so many transcripts say “Speaker 1”, the two ways an app can know who spoke, and how to get real names on yours.

A transcript without names is a wall of text. You can't search for what Maya promised, and you can't send it to anyone without first working out who “Speaker 2” was. Whether you get names depends on how your app decides who is speaking, and there are only two ways it can.

Way 1: Names from the meeting app

Zoom, Meet and Teams know who is in the call and whose microphone is live. A note-taker that joins as a bot, or reads the call app's window, or looks at the calendar invite, can copy those names onto the transcript. Granola tags speakers with each participant's display name from the call app, and Wispr Flow's Notetaker takes names from the invite and from who the call app shows as speaking.

This works well in a call where everyone joined from their own laptop under their own name. It gives out when:

Several people share one room and one microphone. The call app sees a single participant called “Boardroom”.
You're recording people in the room with you, with no call at all.
The call app isn't one the note-taker reads. Then you get “Me” and “Them”, or numbered speakers.
Someone dialled in by phone, or joined as “iPhone”.

Way 2: Names from the voice

The other way ignores the call app and listens. The app first separates the recording into voices, which is what gives you “Speaker 1” and “Speaker 2”. If it stops there, you rename them by hand in every transcript: Superwhisper and MacWhisper keep a name within the transcript you typed it in. If the app also remembers how a named voice sounds, the next meeting comes out with the name already on it.

Recognizing a voice works anywhere you can record: any call app, a shared conference room, a conversation across a table. What to ask is where the voice is kept. Some services keep what they learn about a voice in their cloud. Enzo works out who is speaking on your Mac, and the voices it knows stay there.

Getting names with Enzo

Record a meeting. Enzo records your microphone and the sound from your Mac, so it works with Zoom, Meet, Teams, FaceTime, Slack or people in the room.
Open the transcript in Speak. The voices are already separated. Anyone Enzo doesn't know has a neutral label.
Click a speaker's name and say who it was. The whole transcript updates.
Record the next meeting. Enzo names the people it knows without being asked.

Each line carries a name and a time. Click a name to correct it.

Enzo matches by how a voice sounds, never by a name on an invite, and it doesn't guess. A voice it isn't sure of keeps its neutral label until you say who it is. Open someone in People to see every meeting you've had with them.

Getting better names from any app

Name people after the first meeting, while you remember who was who. One naming is enough for an app that learns voices.
Fix a wrong label when you see one. In Enzo, correcting a line updates the transcript.
Expect mistakes where people talk over each other, where two voices sound alike, or where the audio is poor. A speakerphone across the room is harder than a headset. No app is perfect at this.
In a room, put the Mac where it can hear everyone, not beside the loudest person.

A voice is personal data

An app that knows someone by voice is holding something about them. Tell people you're recording, as you would with any recorder, and know where their voices go. With Enzo, recordings, transcripts and the voices it has learned stay on your Mac and are never uploaded or synced. See what runs on your Mac and the privacy policy.

Which should you use?

Every meeting is a video call where each person joins under their own name: names from the meeting app are enough.
You meet in rooms, record interviews, use several call apps, or talk to the same people every week: use an app that learns voices, and check where it keeps them.

More on how Enzo records meetings on the meeting transcription page, and on recording without a bot.