Codebase
0
Indexes code repositories into PostgreSQL with pgvector for production-grade semantic code search tailored for AI coding assistants.
About
The Codebase server provides advanced semantic code search capabilities through a local-first architecture, enabling AI assistants to efficiently understand and navigate codebases. It achieves this by combining tree-sitter AST parsing with vector embeddings, storing them in PostgreSQL with the pgvector extension. Beyond search, it also offers robust task management tools with Git integration, all exposed via the Model Context Protocol (MCP) for seamless interaction with AI clients.
Key Features
- Production-Ready Performance (e.g., 60-second indexing for 10K files, 500ms p95 search latency)
- 0 GitHub stars
- Development Task Management with Git integration
- Semantic Code Search with natural language queries
- Fast Repository Indexing using tree-sitter AST parsing and chunking
- Comprehensive Error Handling, Structured Logging, and Type Safety
Use Cases
- Streamlining development workflow by linking tasks to code changes via Git
- Empowering AI coding assistants with deep codebase understanding
- Facilitating natural language queries for code exploration
Sponsored