CodeGraph
A local-first code-intelligence tool that turns any codebase into a queryable knowledge graph for AI coding agents.
CodeGraph
What is CodeGraph?
How to use CodeGraph?
- 1Install CodeGraph using npx @colbymchenry/codegraph
- 2Run it on your codebase to build the knowledge graph
- 3Connect MCP-enabled AI agents to leverage the graph for coding tasks
CodeGraph Key Features
- Tree-sitter parsing for 20+ programming languages
- MCP server integration for AI agents
- Impact analysis by tracing callers and callees
- 100% local and private, no external services
- Queriable via CLI, MCP, and TypeScript library
- Deterministic extraction from ASTs, no LLM overhead
CodeGraph Use Cases
- Understand complex codebases quickly
- Assess impact of changes before implementation
- Provide rich context to AI coding agents
- Perform code navigation and analysis
CodeGraph Pricing & Free Credits
CodeGraph currently operates on a Free model.
CodeGraph Pros & Cons
Pros
- 100% local and private
- Fast incremental parsing
- Deterministic AST extraction
- Supports 20+ languages
- Integrates with MCP-based AI agents
Cons
- Requires Node.js setup
- Limited to code intelligence (not a general AI feature)
- May require configuration for complex projects
What is CodeGraph best for?
- Developers
- AI coding agent users
- Code reviewers
CodeGraph FAQ
Top free alternatives to CodeGraph
A dual-arm robot platform for physical AI research and deployment, offering high reach, payload, and control rate for real-world manipulation tasks.
OLO Robotics is a browser-based robotics development platform with built-in AI-assisted coding, simulation, and ROS2 tools for faster robot building.
SeaTicket centralizes support issues from GitHub, forums, and emails into one workspace, using AI agents to autonomously resolve issues.
Spotlight by Backplanes reads your Claude Code and Codex sessions to give you session reports that help you get recursively better.
Agentcad is an open-source MCP server and CLI that enables AI coding agents to design, render, and export CAD output directly from their coding environment.
Kimi Work is a desktop AI agent that automates browser tasks, runs Python, orchestrates up to 300 parallel agents, and schedules workflows locally.
Drop the design system into any codebase and your coding agent instantly knows how to build consistent, polished interfaces.