Repo Chat enables users to query GitHub repositories using natural language. It leverages OpenAI embeddings and Supabase with pgvector for efficient semantic search, making it suitable for developers and researchers seeking to understand codebases without manual inspection.

How It Works

The project utilizes OpenAI's embedding models to convert repository code into vector representations. These vectors are stored in Supabase, a PostgreSQL database enhanced with the pgvector extension. When a user asks a question, it's also embedded, and the system retrieves the most semantically similar code chunks from the vector database to generate an answer.

Quick Start & Requirements

• Install dependencies: pip install -r requirements.txt
• Run data loading: python3 load.py
• Run embedding: python3 embed.py
• Run chat interface: python3 main.py
• Prerequisites: OpenAI API key, Supabase account with pgvector extension enabled, GitHub repository URL and branch.

Highlighted Details

• Uses OpenAI embeddings for semantic understanding.
• Leverages Supabase with pgvector for efficient similarity search.
• Supports querying specific branches of a repository.

Maintenance & Community

• Project maintained by mckaywrigley.
• Contact available via Twitter.

Licensing & Compatibility

• License not specified in the README.
• Compatibility with commercial or closed-source projects is undetermined due to the lack of license information.

Limitations & Caveats

The project relies on external services (OpenAI, Supabase) which may incur costs. The README does not specify a license, raising concerns about usage rights and commercial compatibility.