A better conversation starts here.
Set the customer details, then talk to your collections agent.
Ready when you are
Your microphone. Your scenario. A real conversation.
Conversation
LIVE TRANSCRIPTYour conversation will appear here.
Turn-to-turn latency
End of your speech → first agent audio. Readings appear after each reply finishes.
Speak to the agent to collect your first reading. The opening greeting is excluded.
Measured by the agent; browser playback delay is excluded. LLM and TTS timings overlap and do not add up to the response gap. Showing the latest 8 turns.
Estimated call cost
Start a call to see usage from every provider.
No usage reported yet
No usage reported yet
No usage reported yet
No usage reported yet
How the estimate works
Gross usage at public list prices (VoiceLab uses your estimated rate), before included allowances, discounts and taxes. Soniox STT uses open-stream time, including pauses; Soniox TTS estimates tokens from text and generated audio. Gemini uses reported text and audio tokens when available, with estimated usage otherwise. LiveKit covers connection time at the Ship rate; bandwidth is excluded. Cloudflare hosting and local worker costs are excluded.
Tokens and speech usage arrive after processing. Interrupted responses, retries and usage reported after disconnection can make the final invoice differ.
Per-provider rates appear when the agent connects.