AI gateway with self-validating semantic cache for OpenAI, Anthropic, and Google APIs. Reduces LLM API costs by 50%+ through intelligent caching without breaking response quality. Drop-in OpenAI-compatible base URL replacement.