autoagents-graph 0.1.0

The AI Workflow Cross-Platform Engine

The AI Workflow Cross-Platform Engine

English | 简体中文

Table of Contents

• Why Choose Hermes?
• Quick Start
• Architecture
• Contributing
• License

Why Choose Hermes?

Hermes is a revolutionary AI workflow cross-platform engine that allows you to freely convert workflows between different AI platforms through a unified API. Like Hermes delivering messages between the divine and mortal realms, Hermes helps you seamlessly navigate through complex AI ecosystems.

• Zero Learning Curve: Unified API design - learn once, use everywhere
• Type Safety: Complete type validation based on Pydantic, ensuring secure workflow transmission
• Platform Compatibility: Supports mainstream platforms like Dify, Agentify, with continuous expansion
• Intelligent Conversion: Automatic node type recognition and conversion, as precise as divine prophecy

Quick Start

System Requirements

• Python 3.11+
• pip or poetry

Installation & Setup

# 1. Clone the repository
git clone https://github.com/Hehua-Fan/Hermes.git
cd Hermes

# 2. Install dependencies
pip install -r requirements.txt

# 3. Quick experience
cd playground/text2workflow
python test_text2workflow.py

Basic Usage

Hermes provides three main usage patterns:

Text2Workflow - Cross-Platform Converter

from src.Text2Workflow import Text2Workflow
from src.dify import DifyStartState, DifyLLMState, DifyEndState, START, END

# Create Dify platform workflow
workflow = Text2Workflow(
    platform="dify",
    app_name="Smart Assistant"
)

# Add nodes
workflow.add_node(id=START, state=DifyStartState(title="Start"))
workflow.add_node(id="ai", state=DifyLLMState(title="AI Response"))
workflow.add_node(id=END, state=DifyEndState(title="End"))

# Compile workflow
workflow.compile()

FlowGraph - Agentify Native Builder

from src.agentify import FlowGraph, QuestionInputState, AiChatState

# Create Agentify workflow
flow = FlowGraph(
    personal_auth_key="your_key",
    personal_auth_secret="your_secret"
)

# Build intelligent conversation flow
flow.add_node("input", QuestionInputState(inputText=True))
flow.add_node("ai", AiChatState(model="doubao-deepseek-v3"))
flow.add_edge("input", "ai")

# Publish to platform
flow.publish_app("Smart Chat Assistant")

Running Examples

# Test Agentify platform functionality
cd playground/agentify
python test.py

# Test Dify platform integration
cd playground/dify
python test_dify.py

# Test cross-platform conversion
cd playground/text2workflow
python test_text2workflow.py

Architecture

Core Components

Hermes/
├── src/                        # Core source code
│   ├── agentify/              # Agentify platform engine
│   │   ├── FlowGraph.py       # Workflow graph builder
│   │   ├── NodeRegistry.py    # Node registry
│   │   └── types/             # Node type definitions
│   ├── dify/                  # Dify platform adapter
│   │   ├── DifyGraph.py       # Dify workflow builder
│   │   └── DifyTypes.py       # Dify node types
│   └── Text2Workflow.py       # Cross-platform converter
└── playground/                 # Examples and tests
    ├── agentify/              # Agentify platform examples
    ├── dify/                  # Dify platform examples
    └── text2workflow/         # Cross-platform examples

Design Philosophy

• Unified Abstraction: Workflows from different platforms unified as node-edge graph models
• Intelligent Adaptation: Automatic node type recognition and cross-platform conversion
• Modular Design: Each platform independently implemented for easy extension and maintenance
• Type Safety: Complete type system ensuring compile-time error detection

Supported Node Types

Agentify Platform Nodes

• QuestionInputState - User input node
• AiChatState - AI conversation node
• ConfirmReplyState - Confirmation reply node
• KnowledgeSearchState - Knowledge base search node

Dify Platform Nodes

• DifyStartState - Start node
• DifyLLMState - LLM node
• DifyKnowledgeRetrievalState - Knowledge retrieval node
• DifyEndState - End node

Contributing

We welcome community contributions! Please check the contribution guidelines for detailed processes.

Development Workflow

1. Fork this project
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Create a Pull Request

Contribution Types

• Bug fixes
• New feature development
• Documentation improvements
• Test cases
• Platform adapters

License

This project is licensed under the MIT License - see the LICENSE file for details.

If you have any questions or suggestions, please contact us through Issues.