Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@codegraph-mcp Which functions call parseConfig?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Graph-native code intelligence for coding agents.
This project tests a simple hypothesis: an agent should navigate a repository through a compact, deterministic program graph and retrieve full source only for symbols it decides are relevant.
The MVP indexes TypeScript into a function/symbol-level SQLite graph and exposes structural queries over MCP.
TypeScript Compiler API indexing
stable symbol IDs
function/method call edges
symbol read/write/reference edges
SQLite graph store
compact MCP queries
lazy source retrieval by symbol
Related MCP server: CodeMind
npm install
npm run build
node dist/cli.js index /path/to/typescript/project
node dist/cli.js serve /path/to/typescript/projectThe server uses stdio transport, so an MCP host should launch the serve command and communicate over stdin/stdout.
The index database defaults to <project>/.codegraph.sqlite. Pass a third argument to index and serve to use another path.
search_symbols, get_symbol, get_neighbors, get_callers, get_callees, get_references, and get_source.
See docs/DESIGN.md for the experiment and docs/MCP.md for the tool contract.
Early MVP foundation. Direct-edge correctness, path traversal, stale-index detection, fixtures, and the evaluation harness are next.
This server cannot be deployed
Code intelligence for coding agents: semantic, AST, graph, and full-text search. 279+ languages.
Codebase intelligence for AI agents — dead code, blast radius, ownership.
Code intelligence for LLMs. Analyze, search, and retrieve code from any public git repository.
Shared memory for coding agents. Stop re-explaining your codebase every session.
