🧠💬 GitHub Codebase Assistant

A chat-based AI assistant for deeply understanding and navigating GitHub repositories.

---

🚀 What this Project Does

This tool allows developers to ask questions about any codebase and get reliable, context-aware answers, grounded only in the repository's source files. It’s built on a Retrieval-Augmented Generation (RAG) architecture to eliminate hallucinations and provide high-fidelity technical insights.

✨ Core Features

• Clone & Index: Clones any public GitHub repository for processing.
• Deep Understanding: Parses and chunks source code files for deep, granular comprehension.
• Vectorized Knowledge: Generates embeddings for code snippets and stores them in a vector database.
• Precision Retrieval: Retrieves the most relevant code snippets for every question asked.
• Context-Aware Chat: Interact with the codebase as a conversation.
• Grounded Responses: Responses are guaranteed to be based only on the provided source code.
• Exact References: Get file references (src/path/to/file.js) alongside every answer.

---

🧩 Tech Stack

This project is built using a modern, efficient, and fully self-hosted stack to deliver fast, reliable code insights.

Category	Key Technologies	Description
Frontend	React, CSS/Tailwind	Minimalist, chat-style UI with real-time typing indicators for a responsive feel.
Backend	FastAPI	High-performance Python REST API handling repo loading, indexing, and Q/A requests.
Caching	In-Memory Stores	Used for caching loaded repositories and vector stores to minimize re-indexing time.

🤖 AI / RAG Components

• Orchestration: LangChain
  • Used for managing the RAG pipeline, chaining components, and structured prompting.
• Vector Database: Chroma
  • The persistent, in-memory store for vectorized code embeddings.
• Local LLM Host: Ollama
  • Enables local, private execution of the Language Model.
  • LLM: Llama 3 (The large language model used for generating answers.)
  • Embeddings: mxbai-embed-large (The embedding model used to create vector representations of the code.)

---

❓ Example Questions You Can Ask

• Which function handles login?
• Where is authentication implemented?
• How does this service work internally?
• Explain this class like I’m new to the project.
• What is the purpose of the 'utility.py' file?

---

🛠️ Architecture Overview

This project uses a RAG (Retrieval-Augmented Generation) architecture to deeply understand real codebases instead of hallucinating.