🚀 Welcome to agnflow¶

⚡ Efficient Python Agent Workflow Engine

agnflow is a lightweight, high-performance Python library for building intelligent agent workflows. With its minimalist syntax and powerful features, you can create complex agent systems in just a few lines of code.

🎯 Key Features¶

⚡ Minimalist Syntax¶

Build agent workflows in 5 lines of code with intuitive operators.

🎨 Auto Visual Flowcharts¶

Generate beautiful flowcharts automatically with one line of code.

🔄 Runtime Dynamic Management¶

Add or remove nodes at runtime with symmetric connection/disconnection syntax.

🚀 Advanced Flow Control¶

Support sync/async mixed execution, branching, looping, and swarm agents.

🚀 Quick Start¶

from agnflow import Node, Flow

# Define nodes
greet = Node("Greet", exec=lambda state: {"message": "Hello!"})
respond = Node("Respond", exec=lambda state: print(state["message"]))

# Build and run workflow
flow = Flow(greet >> respond)
flow.run({"data": "hello"})

📚 Learning Path¶

• 🚀 Getting Started - Learn the basics in minutes
• 🧠 Core Concepts - Understand the fundamental concepts
• �� API Reference - Complete API documentation
• 💡 Examples - Ready-to-run examples

🎯 Why agnflow?¶

• ⚡ Lightweight: Core code only hundreds of lines
• 🎨 Visual: Auto-generate beautiful flowcharts
• 🔄 Dynamic: Add/remove nodes at runtime
• 🤖 Agent-Friendly: Native LLM integration support
• 🚀 Fast: Minimal overhead, maximum performance