OPEN-SOURCE VOICE INTELLIGENCE / v1.0
Transcribe natural speech. Create expressive audio. Explore SonicAgent's voice API in your browser—powered by Whisper & Piper.
01 / LISTEN
Record or drop audio. Whisper transcribes it as soon as the recording is ready, without a second click.
Press the microphone. Transcription starts when you stop.
Your transcript will appear here after processing audio.
02 / SPEAK
Give your words a voice. Companion is warmer and slower. Secretary is clearer and a little faster. Voices stay in the language you select, grouped so you can compare them.
Piper is optional and may be disabled in this Space. Readiness is checked automatically.
Connect SonicAgent to a WebXR avatar, an application, or a workflow with standard HTTP endpoints.
Explore API docs →/health/ready/v1/audio/transcriptions/v1/audio/speech/v1/audio/capabilities