Your codebase never sleeps.

Lucid is an open-source AI agent that watches your code, thinks while you're away, and remembers everything.

$ npm install -g @lucid-agent/cli
lucid — ~/my-project

Every AI coding tool forgets everything between sessions. You spend the first ten minutes re-explaining your project, your conventions, your ongoing work. Every single time.

Lucid is different. It runs a lightweight daemon that watches your codebase and decides, every 60 seconds, whether something worth thinking about has changed. Most checks cost less than a tenth of a cent. When it matters, a full analysis runs — code review, pattern detection, memory consolidation — and the findings persist across sessions in a 3-layer memory system that actually grows with your project.

Bring your own model. Claude, GPT, Gemini, or run entirely local with Ollama. Open source, model-agnostic, and designed to stay out of the way until it has something worth saying.

<$0.001
Per tick. Background monitoring costs almost nothing. Real analysis only runs when it matters.
3-layer memory
LUCID.md index, topic deep-dives, and a full SQLite archive with search. Nothing is forgotten.
4 providers
Anthropic, OpenAI, Google, or Ollama for fully local, zero-cost operation. Your keys, your choice.

Coming soon: Lucid Cloud

Don't want to run it locally? We're building a hosted version with GitHub integration, team memory, and a web dashboard.

Thanks! We'll be in touch.