Speaker Diarization
Speaker diarization automatically identifies "who spoke when" in your meetings, making it easy to follow conversations with multiple participants.
What is Speaker Diarization?
Instead of a plain transcript like this:
Hello everyone, thanks for joining. Let me share the quarterly update. That sounds great. I have a question about the budget.
You get a labeled transcript like this:
Speaker 1: Hello everyone, thanks for joining.
Speaker 1: Let me share the quarterly update.
Speaker 2: That sounds great.
Speaker 2: I have a question about the budget.
How to Enable Speaker ID
- 1Go to Settings → Transcription
- 2Toggle "Speaker Identification" to ON
- 3Start a new transcription - speakers will be automatically detected
Speaker Limits by Plan
| Plan | Max Speakers | Speaker ID Available |
|---|---|---|
| Free | 2 speakers | ✗ Not available |
| Basic | 5 speakers | ✓ Available |
| Pro | 10 speakers | ✓ Available |
| Enterprise | Unlimited | ✓ Available |
Editing Speaker Names
After transcription, you can replace generic labels (Speaker 1, Speaker 2) with actual names:
- 1Click on any speaker label in the transcript (e.g., "Speaker 1")
- 2Type the person's actual name
- 3Choose "Apply to all" to update all segments from that speaker
💡 Tip: Speaker names are saved with the transcript. You can also edit them when exporting to include proper attribution.
Tips for Better Speaker Detection
- Avoid overlapping speech - When multiple people talk at once, speaker detection becomes less accurate
- Use quality microphones - Better audio quality leads to better speaker separation
- Speak in turns - Brief pauses between speakers help the AI distinguish voices
- Similar voices - The AI may occasionally confuse similar-sounding voices; manual correction is easy