llm-review

A simple tool to automate GitLab Merge Request reviews using Large Language Models.

Setup Instructions

1. Configuration

Create a file named .env in the root directory of the project.

2. Environment Variables

Add your GitLab Personal Access Token and the link to the Merge Request you wish to review.

.env sample:

GITLAB_TOKEN=your-token
MR_LINK="
https://git.something.com/some-repo/some-project/-/merge_requests/123
"

3. Model downloading (new)

Version 0.1.0 used Qwen2.5 14b and have usefull, but lots of false positive results. Model was downloaded with transformers automatically. Model require ~12GB of VRAM.

Version 1.0.0 use Qwen3.5-35B-A3B and have production ready results. Model require manual downloading, put it into ./llm-review/models/Qwen3.5-35B-A3B-Q5_K_M.gguf. This model can be partially swapped to RAM and work with ~20 tokens/second.

Version 1.1.0 add API support for local models, tested with llama.cpp server and gemini 4 model.

4. Use api

Although this tool is designed to work directly with a local LLM, you may find it more convenient to run the model on a separate server via an API.

To enable API mode, add the following lines to your .env file:

USE_API=true
API_PORT=8000

Usage

Run the script using the following command in your terminal python run.py