Practical overview
When it fits
Use it when you want to study or experiment with LLM agents negotiating and making decisions in Diplomacy, or inspect their recorded game interactions.
Who it is for
- Developers exploring multi-agent language-model behaviour in a turn-based strategy game.
- Researchers or evaluators comparing model decision-making and negotiation in Diplomacy.
- People analyzing logged game interactions, orders and agent relationships.
Documented uses
- Run Diplomacy games with language-model agents representing the powers.
- Experiment with negotiation, relationship tracking and strategic order generation.
- Compare configured models through game outputs and interaction logs.
- Process logged LLM interactions for reinforcement-learning analysis.
Requirements & limits
- The project is a Python repository managed with uv; it is not presented as a hosted game service.
- LLM use requires provider API keys and model assignments in project configuration.
- It supports multiple named model providers, and models can be assigned by game power.
- The README describes a React web interface as a separate component for play and visualization.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up the Python project and configure model providers
Prepare the repository environment and configure API keys and model assignments for its LLM agents.
Before you start
- Python project environment managed with uv
- API keys for the model providers you plan to configure
Steps
Clone the repository and change into its directory.
Run uv sync to install dependencies and create the virtual environment.
Activate the environment using the documented Unix/macOS or Windows command.
Create a .env file with the relevant provider API keys.
Assign models to powers in ai_diplomacy/utils.py.
Expected outcome
The repository environment is set up and provider credentials and model assignments are configured for the agents.
Guide limitations
- These instructions cover setup and configuration, not running or validating an LLM game.
- Provider account, key-creation and model-availability details are not included in the setup instructions.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 7 November 2026
- Original website title
- GoodStartLabs/AI_Diplomacy · Official homepage title · retained in its source language