Generates a function-level dependency graph across multiple programming languages, providing deep code intelligence for AI agents, developers, and CI/CD pipelines.
Sponsored
AI agents frequently struggle with large codebases, consuming excessive tokens to understand structure or making flawed assumptions, leading to broken code and repetitive review cycles. Codegraph resolves this by building and maintaining a function-level dependency graph of your entire codebase, leveraging tree-sitter and SQLite. It exposes this intelligence via an MCP server for AI agents, a CLI for developers, CI gates for quality enforcement, and a programmatic API, ensuring AI agents have structural context before making changes and enabling developers to catch issues proactively.
Características Principales
0136 GitHub stars
0233-tool MCP server for AI agents
03Git-aware impact analysis with co-change detection
04Function-level dependency graph across 11 languages
05Hybrid semantic search (BM25 + embeddings)
06Architecture boundary enforcement and CI quality gates
Casos de Uso
01Enforce architectural rules and quality standards directly within CI/CD pipelines to prevent structural degradation.
02Improve code review efficiency by proactively identifying structural issues before PRs are opened.
03Empower AI agents with comprehensive codebase context for accurate and structurally sound code modifications.