⚠️ Note: Lux is currently under heavy development and should be considered pre-alpha software. The API and architecture are subject to significant changes. We welcome feedback and contributions.
Lux is a powerful language-agnostic framework for building intelligent, adaptive, and collaborative multi-agent systems. It enables autonomous entities (Agents) to communicate, learn, and execute complex workflows while continuously improving through reflection.
- 🧠 Self-Improving Agents: Agents with built-in reflection capabilities (coming soon)
- 🚀 Language Agnostic: Build agents in your favorite programming language
- 🔄 Type-Safe Communication: Structured data flow with schema validation
- 🤖 AI-First: Deep LLM integration with advanced prompting and context management
- 🔌 Extensible: Easy integration with external services and APIs
- 📊 Observable: Built-in monitoring, metrics, and debugging tools
- 🧪 Testable: Comprehensive testing utilities for deterministic agent behavior
📚 Read the full documentation on hexdocs.pm/lux
- Getting Started Guide (docs coming soon) - Start here if you're new to Lux
- Core Concepts (docs coming soon) - Learn about Agents, Signals, Prisms, and Beams
- Language Support (docs coming soon) - Language integration details
- Agents - Building intelligent autonomous agents
- Signals - Type-safe communication between agents
- Prisms - Modular functional components
- Beams - Workflow orchestration
- Lenses - External service integration
- Multi-Agent Collaboration - Build collaborative systems
- Trading System - Complete crypto trading example
- Running a Company - Multi-agent content creation pipeline
- Role Management - Managing agent roles
- Companies - Organizing agents into companies
- Contributing Guide - Help improve Lux
- Testing Guide - Testing your Lux applications
- Troubleshooting - Common issues and solutions
Autonomous agents that combine intelligence and execution. Agents can:
- Monitor and analyze data
- Make strategic decisions
- Delegate tasks to other agents
- Adapt to changing conditions
- Collaborate through structured protocols
Type-safe communication using predefined schemas. Signals provide:
- Structured data validation
- Type safety across language boundaries
- Clear communication protocols
- Versioning and compatibility
Pure functional components for specific tasks. Prisms enable:
- Modular functionality
- Language-specific implementations
- Clear input/output contracts
- Easy testing and validation
Composable workflow orchestrators. Beams allow you to:
- Define complex workflows
- Coordinate multiple agents
- Handle parallel execution
- Manage state and dependencies
Lux provides first-class support for multiple programming languages:
- Python: Deep integration with Python's scientific and ML ecosystem
- JavaScript/TypeScript: Frontend and Node.js support
- Other Languages: Language-agnostic protocols for easy integration
Learn more about language support
Check out these examples to see Lux in action:
- Trading System: A complete crypto trading system
- Content Creation: Multi-agent content creation pipeline
- Research Assistant: Collaborative research system
We welcome contributions! Whether you want to add support for a new language, improve documentation, or fix bugs, check out our Contributing Guide.
- 💬 Discord Community
- 📝 Blog
Lux is released under the MIT License. See LICENSE for details.