claude-context

Why we included this project

Most coding agents only see the files you paste into the prompt, which leaves the rest of the repository invisible to them, and that's where real-world mistakes tend to come from. Claude Context fixes this by indexing the whole codebase into a vector store and exposing it as an MCP server, so tools like Claude Code, Cursor, or Gemini CLI can pull relevant code into context on demand. Semantic search is what makes it useful on large, unfamiliar codebases: it finds code by meaning rather than exact string matches, which helps when you don't know the precise file or symbol to search for. That approach also keeps API costs down, because you load only the pieces the model actually needs instead of entire directories on every request. Developers and small teams running agentic coding tools day to day will get the most value out of it.