• 🚀 WASP
  • 📖 Table of Contents
  • 📖 Overview
    • 🌟 Features
  • 🛠 Prerequisites
  • 🚀 Getting Started
    • Option 1: Quick Start with NPX
    • Option 2: Manual Setup
  • 🔧 Configuration
    • Model Configuration
    • Response Formats
      • UOMI Format
      • OpenAI Format
  • 💡 Usage Examples
    • Interactive Mode
    • Development
      • Custom Model Integration
  • 📊 Performance Monitoring
  • 🔐 Security
  • 🐛 Debugging
  • 📚 API Reference
    • Host Functions
  • Compiled WASM
  • 🤝 Contributing

This development environment allows you to create, test, and debug UOMI agents using WebAssembly (WASM) and Rust. The environment provides seamless integration with both UOMI and third-party LLM services, supporting multiple model configurations and API formats.

• 🔄 Hot-reloading development environment
• 📝 Interactive console for testing
• 🐛 Built-in debugging capabilities
• 🔍 Response analysis tools
• 💾 Conversation history management
• 🔌 Support for multiple LLM providers
• 🔑 Secure API key management
• 📊 Performance metrics tracking

Before you begin, ensure you have the following installed:

• Rust (latest stable version)
• Node.js (v14 or higher)
• WebAssembly target: rustup target add wasm32-unknown-unknown

# Create a new UOMI agent project
npx wasp create

git clone https://github.com/Uomi-network/uomi-chat-agent-template.git
cd uomi-chat-agent-template/agent
npm install
chmod +x ./bin/build_and_run_host.sh
npm start

The environment supports multiple model configurations through uomi.config.json:

{
  "local_file_path": "path/to/input.txt",
  "api": {
    "timeout_ms": 30000,
    "retry_attempts": 3,
    "headers": {
      "Content-Type": "application/json",
      "Accept": "application/json",
      "User-Agent": "UOMI-Client/1.0"
    }
  },
  "models": {
    "1": {
      "name": "Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4"
    },
    "2": {
      "name": "gpt-3.5-turbo",
      "url": "https://api.openai.com/v1/chat/completions",
      "api_key": "your-api-key-here"
    }
  },
  "ipfs": {
    "gateway": "https://ipfs.io/ipfs",
    "timeout_ms": 10000
  }
}

You can run the node-ai service by following the instruction in this node-ai repository.

By doing that, you don't need to specify a url or api-key in the model's configuration, and you will run the production version of the node-ai service.

If you don't have enough resources to run the node-ai service, you can use a third-party service like openAI, in this case you need to specify the url and api_key in the models configuration.

The environment automatically handles different response formats:

{
  "response": "Hello, how can I help?",
  "time_taken": 1.23,
  "tokens_per_second": 45,
  "total_tokens_generated": 54
}

{
  "choices": [{
    "message": {
      "content": "Hello, how can I help?"
    }
  }],
  "usage": {
    "total_tokens": 150,
    "prompt_tokens": 50,
    "completion_tokens": 100
  }
}

$ npm start
UOMI Development Environment
Type your messages. Use these commands:
/clear - Clear conversation history
/history - Show conversation history
/exit - Exit the program

You: Hello, how are you?
Assistant: Hello! I'm doing well, thank you for asking...

Performance Metrics:
- Time taken: 1.20s
- Tokens/second: 45
- Total tokens: 54

// Add a new model in uomi.config.json
{
  "models": {
    "3": {
      "name": "custom-model",
      "url": "https://api.custom-provider.com/v1/chat",
      "api_key": "your-api-key"
    }
  }
}

The environment provides detailed performance metrics:

• Response time tracking
• Token usage statistics
• Rate limiting information
• Error tracking and retry statistics

• API keys are stored securely in configuration files
• Support for environment variable substitution
• Automatic header management for authentication
• Secure HTTPS communication

Built-in debugging features:

• Detailed WASM logging
• Request/response inspection
• Performance profiling
• Error tracing with retry information

The compiled WASM file after test is located in the host/src/agent_template.wasm directory.

Contributions are welcome! Please feel free to submit a Pull Request.