🤖 LLM Code Review Bot

A lightweight, fully local code reviewer powered by Ollama and Qwen/Mistral.
It parses your Python files, extracts functions, and uses an LLM to suggest improvements — just like a real dev reviewer.

✨ Features

• Extracts Python functions using regex (no AST or tree-sitter required)
• Sends each function to a local LLM (via Ollama) for review
• CLI-based — no frontend required
• Outputs reviews in your terminal
• Completely offline — free & private

🚀 Quickstart

1. Install Ollama and run the model

ollama run qwen:7b-chat

2. Clone and install requirements

git clone https://github.com/ErenErenturk/llm-code-review-bot.git
cd llm-code-review-bot
python -m venv .venv && .venv\Scripts\activate
pip install -r requirements.txt

3. Run the review tool

python review.py path/to/your_script.py

You'll see something like:

[debug] Extracted 3 functions
[debug] LLM response: "This function could benefit from more error handling..."

🛠 Tech Stack

• 🧠 Qwen via Ollama
• 🐍 Python 3.10+
• 🧪 Regex-based function extraction
• 📦 No external APIs

📌 Future Plans

• Batch review folders
• GitHub PR comments
• Save review results to Markdown or HTML

🧠 Example Output

💬 “Consider adding a docstring and improving exception handling for this function.”

Contributions welcome — feedback, PRs, and ideas are all appreciated!