Transcriber
| Field | Description | Default |
|---|---|---|
| Provider | Speech-to-text provider | Deepgram |
| Model | Transcription model | Latest low-latency model |
| Language | Transcription language | English |
| Endpointing | Silence duration before treating speech as final | Short |
Synthesizer
| Field | Description | Default |
|---|---|---|
| Provider | Text-to-speech provider | ElevenLabs |
| Voice | Which voice the agent speaks with | — |
| Model | Speech synthesis model | Latest turbo model |
| Streaming | Stream audio as it’s generated | On |