Agent Pitch: LLM-powered soccer simulator where AI agents evolve strategies between matches

A soccer simulation system in which every player on the field is controlled by an AI agent (a program that makes decisions autonomously). Before each match, a large language model generates each player's decision logic; after the match, the system feeds the match log back to the LLM so strategies improve over time.

Each player runs a decide() callback written in Python, JavaScript, or Rust. All generated code executes inside language-specific sandboxes with no arbitrary host access. The system supports OpenAI, Anthropic, Gemini, DeepSeek, OpenRouter, or a local Ollama instance.

The project includes a browser UI (port 8765), CLI commands for running seasons, cup tournaments, and league tournaments, and is distributed under Apache 2.0 license. It requires Python 3.11+ and an API key for at least one supported LLM provider.