LLM Routing based on use case

[BradKML]

Sometimes it is good to switch from an expensive high-powered model to a more economical model, here are some examples:

• https://github.com/lm-sys/RouteLLM
• https://github.com/SomeOddCodeGuy/WilmerAI
• https://github.com/anyscale/llm-router

[phernandez]

I would definitely love to add in multiple model support and perhaps some agentic workfllow ability to basic-memory.

[BradKML]

Some of the basic agentic workflow should probably be their own issue/feature #117