Projects
Things I build and tinker with, mostly around voice AI and agentic coding.
Pipecat tooling
Tools for building, sizing, and running Pipecat voice agents.
- pipecat-context-hub CLI and MCP server that helps coding agents build Pipecat apps faster, using 5x fewer tokens. GitHub ↗
- Tinkerbox Capacity planning for Pipecat deployments. Site ↗
- pipecat-subagents-lab Async voice agents. GitHub ↗
- pipecat-local-stt-server Local WebSocket transcription server with pluggable ASR backends. GitHub ↗
- pipecat-local-tts-server Local WebSocket text-to-speech server with pluggable backends. GitHub ↗
- pipecat-cloud-daily-sip-pstn SIP and PSTN dial-in and dial-out for Pipecat Cloud, with FastAPI and Next.js servers. GitHub ↗
- pipecat-cloud-sipp-sw SIPp load tester for testing SIP and Pipecat Cloud scaling. GitHub ↗
Voice AI apps
Applications built on speech models and live data.
- onoats Granola-style transcriber with diarization. Local-first, with open-source recorder. Site ↗Local models ↗GitHub ↗
- gamealerts Speaks football commentary while reading live game stats from ESPN. GitHub ↗Collector ↗
- hotwired.news A Grok-powered bot that scans and publishes Voice AI announcements. Site ↗
- qwen3-tts-clone-and-speak Voice cloning and text-to-speech on Apple Silicon using Qwen3-TTS and Kokoro. GitHub ↗
Developer workflow
Making coding agents more autonomous.
Fitness
Training tools I use myself.
- quietset 12-week strength training tracker, with load and deload weeks. Site ↗