Skip to content
  1. Fork Pipecat, create a Windows-audio feature branch, get it running in the existing virtualenv.
  2. Build a minimal mic → STT → LLM → TTS → speakers pipeline behind clean interfaces.
  3. Add basic configuration (.env/config file) and a tiny demo to exercise the pipeline.
  4. Implement Windows device enumeration/selection and normalize sample rate/channel formats.
  5. Add robust start/stop/shutdown semantics with guaranteed cleanup on exceptions.
  6. Implement recoverable error handling with retries, timeouts, and backoff.
  7. Enable incremental/streaming STT and dispatch partial hypotheses to the LLM.
  8. Enable incremental/streaming TTS with chunked playback and backpressure management.
  9. Add turn-taking (VAD, barge-in, silence detection) and interruption/continuation rules.
  10. Measure end-to-end latency; tune buffers, chunk sizes, threading, and priorities.
  11. Add structured telemetry (logs, metrics, optional traces) with opt-in configuration.
  12. Build a simple launcher (CLI and minimal GUI) for day-to-day use with profiles.
  13. Document setup, run, and troubleshooting; add a one-command dev workflow.
  14. Prepare Windows-specific fixes as focused commits with tests and docs.
  15. Submit upstream PR(s), iterate on feedback, maintain branch until merged.
  16. Rebase regularly against upstream; tag internal releases; minimize delta until acceptance.

Tested on: Raspberry Pi 4 (4GB RAM, 120GB SATA SSD), Debian 12 Bookworm, Python 3.11.2

  1. Create dedicated user sudo adduser pipecat su - pipecat
  2. System preparation sudo apt update sudo apt install python3 python3-venv python3-pip git build-essential
  3. Create project environment mkdir ~/pipecat-src && cd ~/pipecat-src python3 -m venv venv source venv/bin/activate
  4. Clone repository git clone https://github.com/pipecat-ai/pipecat.git .
  5. Install dependencies pip install —upgrade pip pip install -r pyproject.toml pip install —upgrade build
  6. Build wheel python -m build
  7. Install wheel cd dist pip install pipecat_ai-0.0.92.dev12-py3-none-any.whl
  8. Verify python -c “import pipecat; print(pipecat.version)“
  • Build failspip install python-dotenv fastapi websockets daily-python then retry python -m build
  • Wheel install fails — Use the full filename: pip install pipecat_ai-<version>-py3-none-any.whl