We'd like to ask for the feature of allowing Chunk to be configured to use a LiteLL AI Gateway for the choice of model. This gives us flexibility in choice, as well as additional overhead protections over directly choosing your built-in model or AWS Bedrock, etc.

Currently, when configuring the tool, we can select one of four model providers:

We'd love an option 5 for LiteLLM Proxy, where we can specify a proxy endpoint and auth.