tirth8205/code-review-graph

Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.

What it solves

AI coding assistants often waste tokens and lose precision by re-reading entire codebases or scanning large chunks of irrelevant files during code reviews. code-review-graph reduces this token waste by providing AI assistants with a precise, structural map of the codebase, ensuring they only read the files that are actually affected by a change.

How it works

The tool uses Tree-sitter to parse repositories into an Abstract Syntax Tree (AST), which is then stored as a graph of nodes (functions, classes, imports) and edges (calls, inheritance, test coverage).

When a file is changed, the system performs a "blast-radius analysis" to trace every caller, dependent, and test that could be affected. This minimal set of files is then provided to the AI assistant via the Model Context Protocol (MCP), allowing the AI to focus only on the relevant context rather than the whole corpus.

Who it’s for

  • Developers using AI coding tools (such as Cursor, Claude Code, Windsurf, Zed, etc.) who want to reduce token costs and improve review accuracy.
  • DevOps/CI Engineers looking to integrate risk-scored PR reviews into their GitHub Actions pipeline to identify affected execution flows and test gaps.

Highlights

  • Massive Token Reduction: Achieves a median per-question token reduction of ~65x compared to whole-corpus baselines.
  • Cros-Platform Integration: Auto-detects and configures MCP for a wide array of platforms including Cursor, Zed, Continue, and GitHub Copilot.
  • Broad Language Support: Supports a vast range of languages (Python, JS/TS, Go, Rust, Java, C++, etc.) and Jupyter/Databricks notebooks.
  • Incremental Updates: Re-indexes only changed files, allowing updates on large projects to happen in seconds.
  • Local-First CI: The GitHub Action builds and queries the graph entirely on the CI runner, ensuring source code never leaves the environment.

Related

  • Project
    jahala/tilth

    A structural code navigation tool and MCP server that helps AI agents read codebases more efficiently by providing outlines, definition-first search, and call-chain tracing.

  • Project
    Houseofmvps/codesight

    A zero-dependency tool that pre-compiles codebase structure and documentation into compact context maps and wikis to drastically reduce token consumption for AI coding assistants.

  • Project
    colbymchenry/codegraph

    A semantic code intelligence platform that provides AI agents with a pre-built knowledge graph of a codebase to eliminate slow file-by-file discovery and reduce token costs.

  • Project
    aovestdipaperino/tokensave

    TokenSave is a local MCP server that builds a semantic knowledge graph of a codebase, letting AI coding agents query symbols, callers, and impact instead of scanning files with grep and read, cutting token usage and tool calls.

  • Project
    vitali87/code-graph-rag

    Code‑Graph‑RAG parses multi‑language codebases with Tree‑sitter, stores the AST‑derived structure in a Memgraph knowledge graph, and lets you query, edit and optimise the code via natural‑language prompts powered by an LLM. It supports dozens of languages, runtime trace overlay, and can run as an MCP server for tool integration.