rag-rat repo intelligence MCP server

rag-rat is a helper that reads your code repository and builds a searchable map of it. Once it is connected, your AI assistant can look up where things live in your code, who calls what, and what past decisions were made. It is handy for anyone who works in a codebase and wants their AI to stop guessing.

What this MCP server does

You ask your AI something about your codebase, like where a setting is handled or what calls a function. Your AI sends that question to rag-rat. rag-rat searches the source files, the code graph, the git history, and any notes previous sessions saved. It sends back an answer with the exact files, lines, and confidence labels, and your AI shows you that in the chat.

Click to zoom

What you can do with it

• Find where a piece of logic lives in your code
• See who calls a function and what it calls
• Check which tests touch a file before you change it
• Look up recent commits and past decisions about a symbol
• Spot when new code duplicates something already in the repo
• Save a note about an important rule so future sessions remember it
• List the most important symbols in the repository

Try asking your AI

• “Where is config reload handled?”
• “Show me the callers and tests for parse_config before I edit it.”
• “Does this helper duplicate anything already in the codebase?”
• “Record an invariant on parse_config: reload must not allocate after the scheduler starts.”

What it gives back to you

You get answers in the chat with the source files and line numbers behind them. You also get lists of callers, callees, tests, recent commits, and saved notes, each with a confidence label. When you save a memory, it is stored and shows up again in later answers about the same symbol or path.

Before you start

Who it's for

Developers and anyone who works inside a code repository and wants their AI assistant to understand it.