Why we built RunBot
CI became the choke point of agentic coding. GitHub's UI wasn't built for it. Local runners were unmanageable. So we built the smallest thing that could fix both.
New actions, releases, and what's shipping next.
CI became the choke point of agentic coding. GitHub's UI wasn't built for it. Local runners were unmanageable. So we built the smallest thing that could fix both.
We run AI code review on every PR, every push, automatically. No cloud API, no per-token billing. Here's the architecture and why the economics get interesting fast.
Read more →We distribute RunBot outside the App Store with no Developer ID, no notarization, and no Gatekeeper ceremonies — just ad-hoc signing, Ed25519 verification, and a fully public CI pipeline.
Read more →Cloud CI made sense before AI-assisted development. At 100–300 commits a day, local runners are 5× faster, nearly free, and the only practical way to run serious AI reviews on every commit.
Read more →