Open-source starters I build and maintain
I open-source the scaffolding I wish I’d had: small, production-shaped starters that run in under a minute and show how I build agents and voice AI for clients. MIT-licensed — use them freely.
LangGraph Agent Starter
A production-shaped LangGraph starter: typed state, a plan → act → verify loop, a human-in-the-loop approval checkpoint, and an evals harness — the parts that make an agent reliable, not just demo-able. Runs offline in under 60 seconds.
- ▹Typed agent state
- ▹plan → act → verify loop
- ▹Human-in-the-loop checkpoint
- ▹Evals harness + zero-dep demo
Voice AI Agent Starter
A latency-aware voice AI boilerplate: the STT → LLM → TTS turn pipeline with a per-stage latency budget, barge-in-ready turn handling, and reference Twilio + Retell webhook servers. Simulates a call offline with no keys.
- ▹STT → LLM → TTS pipeline
- ▹Per-stage latency budget
- ▹Barge-in ready
- ▹Twilio + Retell servers
Each starter ships the same discipline I bring to client work — typed state, evals, latency budgets, human-in-the-loop where it matters. Found them useful, or want one adapted to your stack? Book a free 30-minute scoping call →
Let's see if I can take this off your plate
Tell me what you want to automate. On a free 30-minute call I’ll tell you straight whether it’s worth building, roughly what it costs, and how I’d approach it — no pitch, no obligation.
Book my free 30-min AI scoping call →