The Infinite Evolution Skill Engine for AI Agents
An AI Agent Skill Platform featuring Skill Creator, Skill Evolution, Skills Hub, OctoTrace, Skill Wiki, and AutoSkill — empowering AI Agents to learn, adapt, evolve, and accumulate knowledge continuously.
Octopai is a revolutionary AI Agent Skills Exploration, Extension, and Evolution Intelligence Engine built on a powerful core principle: Everything can be a Skill • Skills evolve through continuous learning • Elevating AI Agent cognition.
Transform any resource—web pages, documents, videos, code, datasets, and more—into structured, reusable Skill content. Through intelligent learning and continuous self-evolution, Skills grow and improve over time.
Mission: Explore, Extend, Evolve AI Agent Cognition
- Explore vast knowledge on the internet and various file formats
- Extend AI Agent capabilities through structured, reusable skills
- Evolve skills through intelligent reflection and optimization
Principles: Everything can be a Skill, Skills evolve through learning
- Everything can be a Skill: Transform any resource into structured, AI-ready Skills
- Skills evolve through learning: Every Skill continuously learns from usage, feedback, and interactions, growing more powerful over time
Create skills from any resource:
- One-click URL conversion: Transform URLs into structured Markdown
- Multi-format parser: PDF, DOC, XLSX, PPTX, images, videos, etc.
- Skill templates: Pre-built templates for quick creation
- Quality evaluation: Intelligent quality scoring and optimization
Advanced Feedback Descent algorithm:
- Frontier management: Maintain top-performing skill variants
- Selection strategies: Best, round-robin, weighted, random
- Evolution modes: Skill-only, prompt-only, hybrid
- Real-time tracing: OctoTrace performance monitoring
Comprehensive skill management center:
- Skill repository: Organize, search, and manage skills
- Version control: Track skill evolution history
- Publishing workflow: Draft → Review → Published → Archived
- Marketplace: Share and discover skills
- Semantic search: Intelligent search and discovery
Performance and cost tracing:
- Real-time monitoring: Track skill performance metrics
- Cost tracking: Monitor API costs and usage
- Visualization: Interactive performance dashboards
Knowledge management system:
- Knowledge graph: Connect and concepts
- Pattern recognition: Identify success/failure patterns
- Experience distillation: Extract and generalize from interactions
- Versioned documentation: Track knowledge evolution
Autonomous research and optimization:
- Auto-experimentation: Self-drive skill experiments
- Auto-optimization: Continuous improvement loop
- Discovery management: Track findings and insights
- Python 3.8 or higher
- OpenRouter API key (get one at openrouter.ai)
# Clone the repository
git clone https://github.com/Yuan-ManX/octopai.git
cd octopai
# Install dependencies
pip install -r requirements.txt
# Configure API keys
cp .env.example .env
# Edit .env with your API keysfrom octopai import Octopai
# Initialize Octopai
octopai = Octopai()
# Convert URL to skill
content = octopai.convert_url("https://example.com")
# Create a skill
skill = octopai.create_skill(
name="Data Analyzer",
description="Analyze CSV data files",
content=content
)
# Start evolution
evolution = octopai.start_evolution(
skill_id=skill.id,
mode="hybrid"
)# Convert URL to Markdown
octopai convert https://example.com -o output.md
# Create a skill
octopai create "A CSV analysis skill" -n csv-analyzer
# Start evolution
octopai evolve csv-analyzer# Start backend (runs on http://localhost:8000)
cd web/backend
python main.py
# Start frontend (runs on http://localhost:5173)
cd web/frontend
npm install
npm run devComprehensive documentation available:
Quick links:
octopai/
├── octopai/ # Core logic
│ ├── agents/ # Agent & evolution
│ ├── skills/ # Skill creation
│ ├── evolution/ # Evolution engine
│ ├── tracing/ # OctoTrace
│ └── utils/ # Utilities
├── web/ # Web application
│ ├── backend/ # FastAPI backend
│ └── frontend/ # React frontend
├── docs/ # Documentation
└── examples/ # Usage examples
MIT License - see LICENSE for details.
Contributions welcome! See guidelines (coming soon).
If you like this project, please ⭐ star the repo! Your support helps us grow!
- Issues: GitHub Issues
- Documentation: docs/
Octopai - Empowering AI Agents to Explore, Extend, and Evolve their cognitive capabilities. 🐙✨