About
I’ve been frustrated trying to get the full context when working across multiple repos in AI assisted software development.
So over a weekend, I built h‑codex – a tool that can pull context from scattered repos into Cursor / Claude Code (integrated via MCP); ensuring that they’ve got the full picture when doing the plans/implementation.
How it works:
- Indexes the repos
- Chunks code with AST for optimal chunk boundaries
- Generates embeddings and stores them in pgvector for fast semantic search
✨ Features
- AST-Based Chunking: Intelligent code parsing using Abstract Syntax Trees for optimal chunk boundaries
- Embedding & Semantic Search: Using OpenAI's
text-embedding-3-smallmodel (support forvoyage-code-3planned) - Vector Database: PostgreSQL with pgvector extension for efficient similarity search
- Multi-Language Support: TypeScript, JavaScript, and extensible for other languages
- Multi-Project Support: Index and search multiple projects
- MCP Integration: Seamlessly connects with AI coding assistants through Model Context Protocol