Simulation Environment for Agentic Marketplaces

---

demo.mp4

Magentic Marketplace is a Python SDK for running simulations of agentic markets. You can configure business and customer agents that transact and then run simulations that evaluate the market's welfare.

Learn more about Magentic Marketplace at our documentation website.

Quick Start

1. Configure your environment

   # Clone the repo
   git clone https://github.com/microsoft/multi-agent-marketplace.git
   cd multi-agent-marketplace
   
   # Install dependencies with `uv`. Install from https://docs.astral.sh/uv/
   uv sync --all-extras
   source .venv/bin/activate
   
   # Configure environment variables in .env. Edit in favorite editor
   cp sample.env .env
   
   # Start the database server
   docker compose up -d

2. Run simulations and analyze the outputs

   # Run an experiment (experiment name is optional)
   magentic-marketplace run data/mexican_3_9 --experiment-name test_exp
   
   # Analyze the results
   magentic-marketplace analyze test_exp

   You can also run experiments from python scripts, see experiments/example.py.

   View more CLI options with magentic-marketplace --help.