OpenRouter or LiteLLM inference provider

[aidando73]

🚀 Describe the new functionality needed

Chances the number of inference providers is beyond our capacity to support.

What if we provide either an inference provider to OpenRouter or via LiteLLM?

💡 Why is this needed? What if we don't build it?

This way users can access the majority of our inference providers through one of these providers, without us having to support every single one - which would become unsustainable and make us the bottleneck of the ecosystem.

Other thoughts

Wdyt?