Clarification on chat API behaviour

[hrflr]

I am a novice user, so I apologize if this is a trivial question:

My understanding is that the official OpenAI API is stateless: each request starts from scratch, and the caller needs to add history in the request. This is in contrast with the /v1/chat/completions API provided by GPT4all, in which chat history is preserved in the "Server Chat" and subsequent requests are influenced by previous ones.

a) is my understanding correct?
b) is there a way to reset/clear the server chat via the http API?

Thank you!

[cebtenzzre]

The model actually seeing the previous messages (as opposed to them just appearing in the chat log) is a regression in v3.5, see #3256.
Sorry about that.