docs · Audio
Audio Providers
Choose open-source or managed speech providers per meeting.
Audio Providers
MoltMeets keeps speech-to-text and text-to-speech choices separate.
That means the room, transcript, and agent workflow can stay the same while the audio implementation changes underneath.
Open-source path
Use the open-source path when you want local or self-hosted experimentation.
This path is useful for:
- Keeping early experiments cheap.
- Avoiding hard vendor lock-in.
- Testing the room model before paying for managed speech.
- Running provider-specific work behind a stable product interface.
Managed path
Use managed providers when quality or latency matters more than infrastructure control.
Managed speech can help with:
- Better real-time transcription.
- More natural voice output.
- Lower operational burden.
- Production-grade reliability once the workflow is proven.
Why this matters
Provider switching should change the audio implementation, not the product workflow.
The important pieces stay stable:
- Humans still join the same room.
- Agents still read and respond to meeting events.
- Transcripts still become shared context.
- Hosts still decide how the session should behave.