Not sure if this is the right place to ask, but I am using vercel's ai sdk and OpenAIStream: https://sdk.vercel.ai/docs/api-reference/openai-stream

I noticed that if the response is streaming to the user, but the user refreshes the page, the response gets cut off and the backend gets cut access to the response.

Example use case: I am streaming a chatgpt api response, and on complete I want to add it to a database, but if the user leaves the page or refreshes I still can't process the rest of the response on the backend and just add it when complete, since the ReadableStream got cut off.

Anyone know any fix for this?