🧠 Local AI Code Reviewer

An intelligent code review assistant that runs locally using LangChain, Ollama, and ChromaDB — optimized for real pull requests and real codebases.

---

🚀 Features

• 🔍 Semantic code search with line-level references (file + line number)
• 💬 Local inference using Ollama + Mistral or Qwen
• 📦 RAG-powered insights using LangChain + ChromaDB
• 🧠 Memory support via LangGraph for contextual conversations
• ✅ PR-aware architecture ready for GitHub integration
• 🧼 Clean codebase following service-oriented architecture

---

🏗️ Architecture Overview

LocalRAG/
├── .env
├── .venv/
├── .gitignore
├── Makefile
├── README.md
├── requirements.txt
├── client.py
├── app/
│   ├── __init__.py
│   ├── api/
│   │   ├── __init__.py
│   │   ├── routers/
│   │   │   ├── __init__.py
│   │   │   ├── chat.py
│   │   │   └── github.py
│   │   └── main.py
│   ├── core/
│   │   ├── __init__.py
│   │   ├── config.py
│   ├── services/
│   │   ├── __init__.py
│   │   ├── github/
│   │   │   ├── __init__.py
│   │   │   └── pr_loader.py
│   │   ├── langchain/
│   │   │   ├── __init__.py
│   │   │   ├── chains.py
│   │   │   └── prompts.py
│   │   └── vectorstore/
│   │       ├── __init__.py
│   │       └── chroma.py
│   └── models/
│       ├── __init__.py
│       └── schemas.py
├── data/
│   └── vectorstore/

Services

• VectorStoreService: Handles indexing and retrieval
• OllamaService: Local LLM client
• GitHubService: Pull request integration
• Memory: LangGraph nodes to manage chat history

---

🛠️ Tech Stack

• LangChain — RAG, memory, embedding abstraction
• ChromaDB — Vector database
• Ollama — Local LLM inference (Mistral, Qwen, etc.)
• LangGraph — Memory-aware multi-step conversations
• FastAPI — Lightweight API server
• Python 3.10+

---

📥 Getting Started

1. Clone & Install

git clone https://github.com/yourusername/local-ai-code-reviewer.git
cd local-ai-code-reviewer
pip install -r requirements.txt

2. Start Ollama

ollama run mistral

Or use:

ollama run qwen:7b

3. Embed Your Repo

make ingest

(Make sure your repo is inside the repos/ folder)

4. Launch the API

make run

5. Ask Questions

python client.py

📍 Referencing Code with File + Line

Each code chunk is enriched with metadata during ingestion:

// File: chat_api.py - Line: 47
@app.post("/ask")
async def ask_question(request: QuestionRequest):

This allows the agent to ground its feedback to the exact place in the code.

🔮 Coming Soon

• Web frontend (Streamlit or FastAPI + React)
• Auto-suggested review comments
• LLM self-evaluation for feedback quality
• Multi-agent code review (naming, perf, security, etc.)

🤝 Contributing

Contributions welcome! Feel free to open issues, PRs, or reach out on LinkedIn.

📜 License

This project is licensed under the MIT License.

📧 Contact

linkedin