Compare / Turn detection
Compare voice-agent turn detection
Choose who owns it.
Supporting evidence
Two approaches to the listening layer.
Built into Omni
Included- Turn detection (endpointing) the moment speech ends
- Grounded KB context retrieved inline
- Included in the Omni voice-agent stack
- No separate product to price and integrate
Build it yourself
Your time + infra- Tune your own VAD / silence thresholds per accent and line
- Stand up + host a vector DB and retrieval service
- Glue transcription, endpointing, and retrieval yourself
- Own the latency, scaling, and on-call
Use Omni for the whole loop, or build each component yourself
Omni ($0.05/min for speech + brain) gives you the whole agent end to end: hearing, turn-taking, grounding, reasoning, and speaking. If you keep your own LLM and TTS, use Hear where English transcription fits and own the remaining turn and retrieval logic in your application. Cue grounding is not active on the serving Hear stream, is unavailable, and has no published launch date.
Sources and implementation
Read the Omni guide ↗ and measurement conditions; test interruptions and end-of-turn behavior on your own calls.
Before you choose
Good questions. Clear answers.
Does Omni include turn detection?
Omni includes turn detection and grounded retrieval inside its voice-agent runtime. Your application connects its knowledge and tools to the managed conversation loop.
Can I buy Cue separately for a Hear stream?
Cue is unavailable. Its reserved grounding fields are not active on the serving Hear stream and have no published launch date. Use Omni for the managed loop, or operate the remaining turn and retrieval logic in your application.
Get an API key. Make a phone agent.
Sandbox keys never bill. Fund a live key when you go to production.