MCP server · Developer tools
repo-graph MCP server
Lets your AI map your codebase and jump straight to the files that matter instead of reading everything.
repo-graph is a helper that builds a map of your codebase: what files and functions exist, how they connect, and where a feature starts and ends. It is for people who work in big or tangled projects, where an AI assistant otherwise wastes time opening file after file trying to find the right one. If you use an AI coding assistant like Claude Code or Cursor, this gives it a shortcut.
What is an MCP server? The 30-second version
On its own, your AI can only chat with you and read whatever you paste or point it at. An MCP server is a small helper program that gives your AI a new skill or a connection to something, in this case to a map of your codebase. Once it is connected, your AI can ask this helper questions like where a feature lives or what a change might affect, and the helper answers from the map. You do not run commands yourself; you just ask your AI in plain words.
What this MCP server does
First, repo-graph scans your project once and builds a lightweight map of the pieces inside it: files, functions, classes, routes, and how they link to each other. When you ask your AI something about your code, the AI asks this helper instead of guessing. The helper looks at the map and returns the smallest set of files and connections that answer your question. Your AI then reads only those files and replies to you. It works across many programming languages in the same project, so a frontend and a backend can be traced together.
Click to zoomWhat you can do with it
- Find which files and functions relate to a feature or a bug
- Trace a flow from a web route all the way down to the database
- See what else might break if you change a particular function
- Spot code that nothing seems to use anymore
- Get a quick overview of what languages and main features a project has
- Read only the files that matter for a task instead of the whole project
Try asking your AI
- “Trace the login flow from the route down to the database”
- “What is the blast radius if I change the checkout function?”
- “Which files do I need to fix the bug where new groups show as closed?”
- “What are the main features in this repo and where does each one start?”
What it gives back to you
You get answers in the chat, not files to download. Usually that means a short list of file names and function names, plus a sentence or two explaining how they connect. Sometimes it is a path from one point to another, like a route leading to a handler and then to a database call. If nothing matches, it tells you that too, so your AI does not keep guessing.
Before you start
What you need
- An AI coding assistant that supports MCP, such as Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, Codex, Gemini CLI, opencode, or Kiro
- The uvx tool installed (a small program runner; the install command uses it)
- A project folder on your computer for it to scan
Good to know
It only reads your code and builds a map, it does not change or delete anything, but the map is stored in a hidden folder in your project and gets rebuilt when you ask it to refresh.
Install it with your AI
Add repo-graph MCP server to your AI, no technical skills needed
You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.
Sign in to get the install prompt
Members get a ready-made prompt that lets the Claude desktop app check repo-graph MCP server, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.
Who it's for
Developers and technical teams working in large, multi-language, or multi-service codebases who use an AI coding assistant.
