Summary

Evotown is an open-source evolution testing platform designed to provide a controlled, reproducible environment for validating the effects of AI evolution engines. It targets researchers and developers working with AI agents, offering a flexible, local economy system that avoids cryptocurrency dependencies, thereby simplifying and standardizing the evaluation of evolutionary algorithms.

How It Works

The platform establishes a sandboxed arena where various evolution engines, including OpenClaw-style stacks, Hermes, or custom harnesses, can be integrated via an HTTP ingest API. Evotown acts as the central system of record for results and scoring. Its core advantage lies in a configurable, local economy system that ensures reproducible agent interactions and evolution dynamics, independent of external financial systems.

Quick Start & Requirements

• Prerequisites: Python 3.10+, Node.js 18+. Docker Desktop or Docker Engine with Compose plugin is recommended for deployment.
• Installation:
  • Docker (Recommended): Copy .env.example to .env, configure API keys and base URLs, then run docker compose up -d --build.
  • Local Dev: Requires running the FastAPI backend (pip install -r requirements.txt, uvicorn main:app) and the React + Phaser frontend (npm install, npm run dev) in separate terminals.
• Configuration: Key settings like LLM API keys and base URLs are managed via the .env file, while economy rules are defined in backend/evotown_config.json.
• Links: Main Arena: http://localhost/arena. Ingest API Docs: docs/en/EVOTOWN-ENGINE-INGEST-V0.1.md.

Highlighted Details

• Supports integration with diverse external evolution engines through a documented HTTP ingest API (v0.1 draft).
• Features a configurable "Jungle Law" economy system with adjustable parameters for initial balance, costs, rewards, and penalties, ensuring reproducible agent interactions.
• Provides a web-based UI accessible at http://localhost, with dedicated routes for the main arena, task history, and generated evolution chronicles.
• The frontend utilizes WebSockets for real-time arena updates, with a REST fallback for broader compatibility.

Maintenance & Community

Information regarding notable contributors, sponsorships, community channels (e.g., Discord, Slack), or roadmap details is not provided in the README.

Licensing & Compatibility

The README does not specify a software license, which prevents an assessment of compatibility for commercial use or closed-source linking.

Limitations & Caveats

The ingest API is noted as a "v0.1 draft," suggesting potential instability or future breaking changes. The platform's core functionality implicitly relies on external Large Language Model (LLM) APIs, which require separate configuration and incur associated costs or dependencies.