Projects

Open-source tools I started building because the same frictions kept appearing in the work.

Software is now being generated faster than humans can reliably review it. That shifts the pressure onto the surrounding infrastructure: trust boundaries, coordination between tools, and the fidelity of architecture decisions under speed. That is the space these three projects live in.

Approval infrastructure for AI actions

Approva

Human approval for risky AI agent actions — with passkey identity, scoped capabilities, and a verifiable audit trail.

When an agent or an automation wants to do something consequential, Approva pauses it, verifies a real human via passkey, issues a scoped capability for that specific action, and records the entire chain. Open Core with a Cloud Beta available now. Self-hostable.

View project Website ↗GitHub ↗

approvalshuman-in-the-loop

Orchestration bridge for coding agents

Codencer

A persistent daemon that manages, executes, validates, and audits tasks performed by external coding agents.

View project Website ↗GitHub ↗

orchestrationplanner-agent

AI-powered infrastructure orchestrator

Rhodd

Turns architecture definitions into production-ready code, infrastructure, and CI/CD pipelines. Zero boilerplate.

View project Website ↗GitHub ↗

deterministic architecturecompiler

Contact

Interested in one of these projects?

Reach out if the conversation is about the product thesis, architecture, execution model, or a shared problem space.

Start with email