Getting Started
Start generating transcripts for your bots' meetings.
Transcription must be enabled
Bots do not generate transcripts by default.
To generate a transcript, a bot must be configured to use a transcription method.
Transcription Methods
There are two methods bots can use to generate transcriptions:
- Meeting Caption Transcription: Using native meeting platform captions
- AI Transcription: Through one of our transcription partners
The table below outlines the pros and cons of each method:
| Transcription Method | Pros | Cons |
|---|---|---|
| Meeting Captions | - Free - Great diarization since it's based on the meeting provider's separate audio streams | - No per-word timestamps - Can be lower quality depending on the platform |
| AI Transcription | - Advanced features such as custom word dictionaries - Can be more accurate | - Costs additional money |
Fetching Transcripts
Regardless of the transcription method you choose, you can fetch a bot's transcript through a single endpoint by calling Get Bot Transcript.
If you're using real-time transcription, you can also leverage Transcription Events to receive partial transcription results in real-time.
Enhanced Diarization
When fetching a bot's transcript, there is an option to enable enhanced_diarization.
This diarization method uses additional punctuation heuristics to improve the accuracy speaker assignment in transcripts.
While this option can improve the accuracy of diarization, we recommend using it only if:
- You are experiencing issues with the default speaker timeline diarization
- You are not using machine diarization
- The transcript language relies on the following punctuation marks:
.,!,?
Enabled enhanced diarization for languages that do not rely on these punctuation marks can result in improper diarization and decreased transcript quality.
Updated 10 months ago