Is it possible to stream the AI's response?

[AFatNiBBa]

I'm using Cheshire Cat with Gemini 2.5 which is very slow when it generates a lot of text, do we have a way of getting the generated tokens a bit a time like ChatGPT does?
This would allow the user to start reading the response before Cheshire Cat has done writing it

[pieroit]

streaming depends on the LLM adapter in langchain, it works with most providers and also with local runners

I guess we need to check the Gemini LLM config and whether it supports streaming

Leaving this issue open as I have no time to check this