Metadata-Version: 2.5
Name: agentship-voice
Version: 0.0.4
Summary: AgentShip voice — run the same agent over a live audio stream. A cascaded pipeline (transport/VAD → STT → agent → TTS) with barge-in, hosted in either Pipecat or LiveKit Agents. The agent is the one the REST service runs; only the hosting differs.
License-Expression: Apache-2.0
Requires-Python: >=3.13
Requires-Dist: agentship-core==0.0.4
Provides-Extra: all
Requires-Dist: livekit-agents<2,>=1.8; extra == 'all'
Requires-Dist: pipecat-ai<2,>=1.8; extra == 'all'
Provides-Extra: livekit
Requires-Dist: livekit-agents<2,>=1.8; extra == 'livekit'
Provides-Extra: pipecat
Requires-Dist: pipecat-ai<2,>=1.8; extra == 'pipecat'
Description-Content-Type: text/markdown

# agentship-voice

Run the same AgentShip agent over a live audio stream.

```bash
pip install "agentship-voice[pipecat]"   # or [livekit]
```

The agent is the one the REST service runs — same spec, memory, tools and tracing. A voice
framework supplies the ears (transport, VAD, STT) and the mouth (TTS); this package supplies
the agent in the shape that framework expects.

See [`docs/capabilities/voice.md`](../../docs/capabilities/voice.md).
