After 3 years of broken AI coding diffs, I built a local-first CLI that verifies its own code — an
Updated 1 hour ago
Project Details
Tired of AI coding agents that hallucinate, break builds, ship corrupt patches, and call it "done." So I spent the last several months building Daedalus a local-first CLI developer workbench designed to run against Ollama/LM Studio (or your own remote API keys) with nothing leaving your machine unless you route it out. 🛡️ Why it's different from standard chat UIs: Self-Verifying Patches: When Daedalus edits code, it diffs TypeScript compiler output (tsc --noEmit) and r