p99-runtime 0.1.0

Local open-source P99 voice runtime (ASR → LLM → TTS over WebSocket)

p99-runtime

Local open-source voice runtime for the P99 protocol.

Speaks the same WebSocket contract as the hosted P99 runtime, so the browser client (@p99labs/voice) works unchanged.

Quick start

pip install p99-runtime
p99-runtime serve --port 8787

Stub mode works with zero models — useful to verify the client pipeline (mic → WS → PCM playback). To run real models, replace handle_turn in p99_runtime/server.py with your ASR → LLM → TTS (or S2S) stack; keep the protocol events and the client stays unchanged.

Endpoints

Path	Description
GET /healthz	Liveness
GET /config	Mic/TTS sample rates for the client
WS /ws	Voice session (JSON control + binary PCM)

The server binds 127.0.0.1 by default. Pass --host 0.0.0.0 only when you have TLS and auth in front (see self-host notes).

Tests

pip install -e '.[dev]'   # from the repo's runtime/ directory
pytest

Hosted runtime (coming soon)

P99 will offer a managed GPU runtime speaking this same protocol — swap the URL, add an API key, no client changes. Watch the repo for availability.