OpenRouter AI Service Skill

功能

• Unified access to 200+ LLMs via OpenRouter
• OpenAI-compatible API interface
• Intelligent model routing and fallbacks
• Real-time streaming responses
• Cost tracking and optimization strategies
• Function calling and vision model support
• Comprehensive error handling and retry logic

使用场景

• Building multi-model AI applications
• Optimizing LLM inference costs
• Implementing model A/B testing
• Avoiding vendor lock-in with diverse model access
• Developing applications requiring real-time LLM responses

非目标

• Managing individual LLM provider accounts or keys
• Providing a direct client for any single LLM
• Abstracting LLM-specific advanced features beyond OpenAI compatibility
• Performing local AI model execution

工作流

1. Obtain OpenRouter API key
2. Install necessary SDKs (e.g., OpenAI SDK)
3. Configure API key as an environment variable
4. Implement API calls using provided code examples for chat, streaming, function calling, etc.
5. Incorporate error handling, retries, and cost optimization strategies
6. Handle responses, including streaming data

实践

• API Integration
• LLM Operations
• Cost Management
• Error Handling
• Security Best Practices

先决条件

• OpenAI SDK (npm/pip)
• OpenRouter API Key
• Environment variable `OPENROUTER_API_KEY` configured