This is the repository for the paper StoryBox: Collaborative Multi-Agent Simulation for Hybrid Bottom-Up Long-Form Story Generation Using Large Language Models, accepted by AAAI 2026.
StoryBox is a framework that leverages collaborative multi-agent simulation for hybrid bottom-up long-form story generation. By combining bottom-up character-driven agent interactions with top-down narrative planning, it dynamically constructs deep, coherent, and engaging story worlds.
We use uv for extremely fast Python package management.
1. Install uv (if you haven't already):
Please refer to the official uv documentation for installation instructions.
2. Clone the repository:
git clone https://github.com/amcghm/StoryBox.git
cd StoryBox3. Install dependencies:
uv syncBefore running the simulation, you need to configure the settings and set up your API key.
The main configuration file is located at reverie/config/config.py. You can adjust parameters such as the LLM model (llm_model_name), temperature, and sandbox settings here.
Setting up the API Key:
The configuration is designed to safely read your API key from environment variables. You can set it in your terminal before running the code:
export OPENAI_API_KEY="your-openai-api-key"Once everything is installed and configured, you can start the StoryBox generation process directly using uv:
cd reverie
uv run run.pyThe output logs, database, and story JSONs will be saved in the output/ directory as specified in the configuration.
If you find our work helpful in your research, please cite our paper:
@inproceedings{chen2026storybox,
title = {StoryBox: Collaborative Multi-Agent Simulation for Hybrid Bottom-Up Long-Form Story Generation Using Large Language Models},
author = {Chen, Zehao and Pan, Rong and Li, Haoran},
booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
volume = {40},
number = {36},
pages = {30359--30367},
year = {2026}
}This project is licensed under the MIT License.