code-graph-mcp
An open-source MCP server for developers using Claude Code–style AI coding agents to generate an AST-based code knowledge graph and produce more traceable semantic search, call-chain analysis, and change-impact insights.
Tool overview
Based on the available evidence, code-graph-mcp is best treated as an early-stage open-source infrastructure component with a clear technical direction, not yet a broadly adopted mature tool. The current evidence is almost entirely the project’s own GitHub repo description, which supports its intended capabilities and positioning, but does not validate adoption, reliability, or outcomes across real projects. In other words, the repo listing is a capability claim, not proof of usability; there is also no visible GitHub growth, hands-on review, tutorial, or long-form field report to strengthen confidence.
Its practical role is not a general code editor, not a standalone IDE, and not an all-purpose autonomous software engineering agent. A more accurate analogy is a “repository structure and relationship analysis backend” attached to an AI coding agent. From the repository text, it uses AST parsing and Tree-sitter indexing across multiple languages to support semantic search, call-graph traversal, HTTP route tracing, and impact analysis.
This tool does not have related social references to display yet.