Speaker identification

AI meeting transcription with speaker identification

note1 knows who said what. Speaker diarization separates the voices, real names come from the meeting roster and the conversation itself, and every speaker label is matched to a real person on your team.

Every identified speaker flows straight into the full meeting transcription report with summaries, action items, and search.

Real names, not Speaker 1
Speaker diarization built in
Remembers every speaker

Transcript

Product review - Google Meet

Speakers identified
M
Maya Lindberg-00:38Roster

Can we confirm what is still blocking the release checklist?

D
Danny Okafor-04:12Matched

The launch date stays the same, but QA ownership needs to be assigned today.

E
Elena Ruiz-09:44Remembered

I will share the customer questions and post the updated checklist after this call.

Every line is attributed to a real person, not an anonymous label.
How it works

Five layers of automatic speaker recognition

No single trick identifies every speaker in every meeting. note1 stacks five mechanisms, from live roster capture to speaker diarization and conversational name detection, so transcripts with multiple speakers come back labeled with real names.

Live roster capture

The note1 note taker joins the call and reads real participant names as each person speaks, so transcripts start with names instead of Speaker 1.

AI speaker diarization

Speaker diarization separates distinct voices in any recording, so multi-speaker audio splits into clean per-speaker segments before names are known.

Conversational name detection

AI reads the conversation itself: self-introductions, direct address, and hand-offs reveal which name belongs to which voice.

Profile matching

Detected names are matched to calendar attendees, teammates, and contacts, including spelling variants and abbreviated names.

Speaker memory

Confirm a speaker once and note1 remembers. The same person is labeled automatically in every future meeting transcript.

Live roster capture

Real names straight from the meeting

When the note1 note taker joins a scheduled Google Meet call, it tracks who is speaking as the meeting happens. The transcript is labeled with actual participant names from the first minute, without diarization guesswork or manual cleanup.

In this meeting

Live
M

Maya Lindberg

Speaking now

Speaking
D

Danny Okafor

In the call

E

Elena Ruiz

In the call

Speaking turns are captured with names attached, so the transcript never starts anonymous.

Voice separation

3 speakers detected
Speaker 1Maya Lindberg
Speaker 2Danny Okafor
Speaker 3Elena Ruiz
Anonymous speaker tracks are upgraded to real names by the layers below.
Speaker diarization

Separate every voice in the recording

For uploads and recordings without roster data, AI speaker diarization detects how many people are talking and splits the audio into per-speaker segments. That speaker separation is what makes it possible to transcribe audio to text with speaker identification instead of one merged wall of words.

Conversational name detection

The conversation reveals who is speaking

People say names constantly: "I'm Maya, I lead design", "thanks, Danny", "Elena, can you take this one?". note1 reads those cues in the transcript and attributes each voice to the name the conversation itself provides, with the exact quote kept as evidence.

Speaker 2 - 00:41

Good point. Thanks, Maya — I will pick that up after standup.

Direct address identifies the previous speaker
M

Speaker 1 identified as Maya

Evidence: “Thanks, Maya” at 00:41

Speaker match

Transcript name matched to a workspace profile

Heard in the meeting

Danny Okafor

Matched to calendar attendees and contacts
D

Danny O.

danny@yourteam.com - Calendar attendee

Spelling variants and short profile names still match the right person.
Profile matching

Names become people on your team

A detected name is only half the answer. note1 matches it against calendar attendees, teammates, and contacts, handling spelling variants and abbreviated profile names, so the speaker connects to a real person with an email, not just a text label.

Speaker memory

Confirm once, recognized in every meeting

When you confirm or correct a speaker, note1 learns the alias. The next time that person joins any meeting in your workspace, their transcript lines are labeled automatically, so speaker identification gets more accurate the more your team uses it.

E

Elena Ruiz

Known speaker in your workspace

Product review

Mon

Confirmed by you

Design sync

Wed

Labeled automatically

Launch retro

Fri

Labeled automatically
Meeting transcription

Identified speakers power the whole meeting report

Speaker identification is not a standalone trick. Every resolved name flows into the searchable meeting transcript, AI summary, and action items, so ownership, decisions, and follow-ups are attributed to the right person automatically.

Use Cases

How teams use an AI note taker

Team Meeting Notes

Review team syncs, planning calls, and weekly updates with AI meeting notes, searchable transcripts, and visible follow-ups.

View use case

AI Meeting Assistant for Managers

Catch up on decisions, review action items with source context, and return to the meeting moments that need your attention.

View use case

AI Interview Notes

Revisit candidate answers, interview transcripts, timestamps, and speaker context after Google Meet interviews end.

View use case

Sales Call Notes

Capture customer requirements, searchable sales call transcripts, and follow-up actions after Google Meet sales calls.

View use case

Product Meeting Notes

Turn product reviews and feedback calls into searchable transcripts, decisions, questions, and follow-up context.

View use case
Platform guides

Meeting notes for the video platforms your team already uses

Focused pages for each meeting platform, without making every page compete for every platform keyword.

FAQ

Frequently Asked Questions

Support
How does AI transcription with speaker identification work?

note1 combines several identification layers. In live meetings the note taker captures real participant names from the meeting roster while each person speaks. For recordings, AI speaker diarization separates the voices first, then conversational name detection and profile matching against your calendar attendees, teammates, and contacts turn anonymous speaker labels into real names.

What is speaker diarization?

Speaker diarization is the AI process of detecting how many people speak in an audio recording and separating the transcript into per-speaker segments. It answers "who spoke when" before any name is known. note1 uses diarization as the base layer of speaker identification whenever roster names are not available.

Can note1 replace Speaker 1 labels with real names?

Yes. Generic labels like Speaker 1 and Speaker 2 are upgraded to real names using the meeting roster, names spoken in the conversation, and matches against your calendar attendees and contacts. Anything note1 cannot resolve automatically is a one-click confirmation, and the correction is remembered for future meetings.

How do I identify speakers in a meeting recording?

Record or upload the meeting to note1. Speaker diarization separates the voices, conversational name detection finds names mentioned in the discussion, and profile matching links them to real people on your team. You can review the suggestions and confirm or adjust any speaker in one click.

Does speaker identification work with multiple speakers?

Yes. note1 transcribes meetings with multiple speakers and keeps every voice separate, so group calls, interviews, and workshops produce a transcript where each contribution is attributed to the right person.