Impact RAG is a repo-aware code analysis tool that uses retrieval-augmented generation (RAG) to answer questions directly from a codebase. Instead of guessing or relying on general knowledge, it indexes a repository using abstract syntax trees, pulls in the most relevant code snippets, and generates answers that stay grounded in the actual code.
The system currently only supports Python repositories with AST-based symbol extraction and will be extended with tree-sitter to support additional languages in the future.
Backend:
- Python
- FastAPI
- LangChain
- Groq
- ChromaDB
Frontend:
- next.js
- tailwind