Evolution / README.md
johangras's picture
๐ŸŒŸ Evolution Aurora - HF Agents MCP Hackathon 2025 Submission
d940355
---
title: Evolution Aurora - AI Learning to Code
emoji: ๐ŸŒŸ
colorFrom: purple
colorTo: green
sdk: gradio
sdk_version: 5.33.1
app_file: app.py
pinned: true
tags:
- track3-agentic-demo
- modal-sponsor
- evolution
- visualization
- ai-agents
- mcp
- hackathon
license: mit
---
# ๐ŸŒŸ Evolution Aurora - Watch AI Learn to Code
**HuggingFace Agents-MCP Hackathon 2025 Submission**
## ๐Ÿ† Track 3: Agentic Demo Showcase
Experience the magic of AI evolution in real-time! Watch as artificial intelligence learns to write better code through evolutionary algorithms, visualized with stunning aurora particle effects.
### โœจ Key Features
- **Aurora Particle Effects**: Beautiful visual bursts triggered by fitness improvements
- **3D Fitness Landscape**: Watch programs climb toward optimal solutions
- **Real-time Evolution**: See AI improve code generation by generation
- **Interactive Controls**: Start/stop evolution and observe the process
### ๐Ÿค– Agent Integration
This demo showcases how AI agents can evolve and improve code autonomously using the [Evolve framework](https://github.com/johan-gras/evolve). The visualization makes the abstract process of machine learning tangible and beautiful.
### ๐Ÿš€ Sponsor: Modal
Designed to leverage Modal's cloud infrastructure for parallel evolution execution (demo runs in simulation mode).
### ๐Ÿ› ๏ธ Technical Stack
- **Frontend**: Gradio with custom HTML5 Canvas
- **Visualization**: Plotly for 3D landscapes
- **Effects**: Pure JavaScript particle system
- **Framework**: Evolve (evolutionary algorithms)
### ๐ŸŽฎ How to Use
1. Click "Start Evolution" to begin
2. Watch the aurora effects burst with each improvement
3. Observe programs climbing the fitness landscape
4. Monitor real-time progress in the event log
Built with โค๏ธ for the HF Agents-MCP Hackathon 2025