Central AI · Voice Engine Latency

Real-Time Call Performance & Streaming Audio Diagnostics

Voice Engine Live
Live Testing Available: Call Leo on WhatsApp at +63 968 749 3955 from your phone (+639267200480) to talk live in streaming Jarvis voice, or use the buttons below to trigger an outbound test call.

Outbound Voice Calling

Engine standby · Ready for interactive voice turn.

End-to-End Latency Streaming
--ms
Caller speech stop → first audio played
LLM Clause Stream First Token
--ms
Central AI native clause generation
TTS Voice Synthesis Jarvis
--ms
Fish Audio real-time chunk synthesis
Speech Recognition (ASR) AU-EN
--ms
whisper-base.en · continuous stream
Call Connection State IDLE
Current TargetEngine Standby
Call Modenative (Jarvis voice)
Answered StateNo
Answered At--
Voice Engine Turns0
Last Spoken Recognition--
Audio Stream & Network Packets Real-Time Metrics
RTP Packets Received0
Decoded Audio Frames0
Played Audio Frames0
Audio Overruns0
Voice Interruptions0
Padding Avoided0 ms
Engine Stream Events Auto-updating every 1.0s
[Init]Connecting to Central AI monitoring endpoint...