AI Generation Client

Robust AI API integration with retry logic and content safety.

When to Use This Skill

• Integrating with AI generation APIs (Gemini, OpenAI, etc.)
• Need retry logic for flaky AI services
• Handling rate limits gracefully
• Detecting content policy violations
• Supporting multi-turn refinements

Core Concepts

AI API integration requires:

1. Exponential backoff - Retry with increasing delays
2. Rate limit handling - Respect Retry-After headers
3. Content safety - Detect and handle policy violations
4. Multi-turn context - Enable cheaper refinements