I would add one more think for RAG being usable in prod. When customer

interacts with language models they expect behavior of streaming text.

Current UX experience is quite bad as user need to wait for whole

response and thinks that something is broken. I would consider this as

very important feature. Is there any workaround, or any change that this

could get into roadmap? https://github.com/typesense/typesense/issues/1727