[](https://github.com/Akshay090/svg-banners)
# ๐๐ป MAGIS ๐๐ป
๐ arXiv Paper ยท ๐ NeurIPS 2024 ยท โ Citation
## โจ Key Features
* ๐ชถ **Simple LLM-based Multi-Agent Framework for Software Evolution**
* ๐ **Visualization of Multi-Agent Processing**
## ๐ฆ Environment
> All commands below have been tested on a Linux-64 machine. The specified package versions are recommended, as they have been verified to work.
```bash
conda create -n magis python=3.11 -y
conda activate magis
conda install redis==5.0.3 -y # If you need to cache repeated requests to the same LLM
pip install openai==0.27.8 pygments==2.17.2 pydeps==1.12.17 tiktoken==0.5.1 redis==5.0.1 flask==3.0.0 datasets==2.16.1
```
## ๐ Run
To use MAGIS on a Git repository, please run the [magis.py](magis.py) script.
You can view options and usage instructions by executing: `python magis.py --help`.
To reproduce our experiments on [SWE-bench](https://github.com/SWE-bench/SWE-bench/tree/b93ac87), please refer to [SWE-bench.md](exp/SWE-bench/README.md).
## ๐ Citation
If you find [MAGIS](https://github.com/co-evolve-lab/magis) useful for your research or applications, feel free to give us a star โญ or cite us using:
```bibtex
@inproceedings{magis,
author = {Wei Tao and
Yucheng Zhou and
Yanlin Wang and
Wenqiang Zhang and
Hongyu Zhang and
Yu Cheng},
title = {{MAGIS:} LLM-Based Multi-Agent Framework for GitHub Issue Resolution},
booktitle = {NeurIPS},
year = {2024}
}
```
## ๐ Acknowledgements
- We thank the prior work, **[SWE-bench](https://arxiv.org/abs/2310.06770)** and [their responses for my question](https://github.com/SWE-bench/SWE-bench/issues/23).
- We gratefully acknowledge [OpenMoji](https://openmoji.org/) for providing the emojis used in our demonstrations.