Noticed the OpenRouter integration in your docs here.

Wanted to flag Anannas as another option in that space; a unified inference layer with 500+ models (OpenAI, Anthropic, Mistral, Gemini, DeepSeek, etc.) with automatic failover, reserved capacity overflow deeper observability — cache analytics, tool-call metrics, model efficiency scoring. Running production workloads; 100k+ requests with stable latency; 4% markup vs OpenRouter's 5.5%, upto 50% faster in some cases.

If you're open to listing alternatives or additional LLM gateway integrations alongside OpenRouter, happy to help with the integration and/docs or example code.

Disclaimer: I'm from Anannas