CodeLore
A versioned knowledge base for your codebase. Capture lessons learned, document system rules, and export context bundles that any developer or AI model can use — without reading every file every time.
What is CodeLore?
CodeLore lets you build a structured, versioned collection of knowledge about your codebase. Each piece of knowledge lives at a path like auth/jwt or api/rate-limiting. Paths form a tree that you can browse in the CLI or web UI, and every time you update a lesson a new version is saved — so you can see how your understanding evolved over time.
CodeLore is designed to feed that knowledge to AI agents. Run codelore context "authentication" to get a clean markdown document of every relevant lesson to paste into a prompt — or connect the built-in MCP server so tools like Claude Desktop and Cursor can search, read, and even write lessons on their own. See the AI Integration guide.
Two modes
🗂 Local mode
Works like git. Run codelore init in any project directory to create a .codelore/ folder with a local SQLite database. No login, no server, works offline. Commit the folder to git or add it to .gitignore — your choice.
☁️ Cloud mode
Sign up, create a workspace, and every lesson you add is stored on the server. Invite teammates to the same workspace. Use the web UI to browse and edit. Use push and pull to sync with a local copy.
codelore init for local work, then codelore login + codelore push to share it with your team.