orientLABSCO SUMMARY
Its own benchmark says where this pays off — and where it doesn't.
On a Go and Angular monorepo of 566 nodes and 620 edges, the same bug fix took 29,838 tokens with repo-graph against 75,308 without, exploring 2 files instead of about 15. The README is equally direct that on a small single-language project the model can just read the code; this earns its keep on large repos, monorepos and multi-service systems.
What it is
A Python server that scans a repository once and builds a graph of entities, relationships and end-to-end flows, then exposes 6 tools the model uses to find the minimum set of files it actually needs.
What you get
orient— the first call on a repo: node and edge counts, detected kinds, entry points, and a blind-spots note flagging which languages or edges are under-linked so you know where to grep insteadfind— turns a symbol, a pasted stacktrace, a failing-test id or a diff into the ranked nodes it implicates, every row carryingpath:lineimpact— blast radius forward or backward as a ranked closure, each row with the edge reason and a marker where the engine finds a node unreachabletrace— one feature end to end with each hop labelled by mechanism (call, HTTP, queue, event) and cross-service hops marked; with two arguments, the shortest path between themread— a node's exact source sliced by its line span, plus a context footer naming HTTP method, cross-stack callers, covering tests and governing docsrefresh— incremental rebuild in which only changed files reparse; routine edits are picked up by the file watcher
Requirements
Run with uvx mcp-repo-graph --repo <YOUR_PATH_TO_PROJECT>, or pip install mcp-repo-graph for a persistent command on your PATH. --repo also accepts a public git URL, which it shallow-clones for you (needs git). uvx mcp-repo-graph install wires up every agent it finds in one command.
Setup effort
One command — uvx mcp-repo-graph install
USE CASES
01
Landing in a codebase you have never seen
Get the counts, the entry points and an honest note about which parts of the graph are thin before you start reading.
02
Turning a stacktrace into the files that matter
Paste the trace or the failing test id and get back ranked nodes with file and line, then read just those.
03
Checking what a change will break
Ask for the forward closure of the functions you're touching, with the edge reason on each row and dead branches marked.
04
Following a feature across a language boundary
Trace from the frontend call to the route and the handler behind it — the link a text search structurally cannot make.
6 tools
Listed in project docs. Your client may show a different list.
findimpacttracereadrefreshBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Requires
uvx on PATH
INSTALLATION
Connect your client
Server setup
TERMINAL
uvx mcp-repo-graph installMCP Clients
Claude Code
CLAUDE CODE · TERMINAL
claude mcp add repo-graph -- uvx mcp-repo-graph --repo .VS Code
VS CODE · TERMINAL
code --add-mcp '{"name":"repo-graph","command":"uvx","args":["mcp-repo-graph","--repo","${workspaceFolder}"]}'VS Code Insiders
VS CODE INSIDERS · URL
https://insiders.vscode.dev/redirect/mcp/install?name=repo-graph&config=%7B%22command%22%3A%22uvx%22%2C%22args%22%3A%5B%22mcp-repo-graph%22%2C%22--repo%22%2C%22.%22%5D%7D&quality=insidersCursor
CURSOR · JSON
{
"mcpServers": {
"repo-graph": {
"command": "uvx",
"args": [
"mcp-repo-graph",
"--repo",
"<YOUR_PATH_TO_PROJECT>"
]
}
}
}
