mirror of
https://github.com/open-webui/open-webui.git
synced 2026-11-17 15:53:59 -06:00
[GH-ISSUE #10686] Feature request: Ability to Continue Response as a new separate message? #135719
Reference in New Issue
Block a user
Originally created by @ferret99gt on GitHub (Feb 24, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/10686
Feature Request
Is your feature request related to a problem? Please describe.
I use LLMs to write long form content, often limiting the response with max_tokens/num_predict to read over, proofread and make edits, before allowing the AI to continue from where it was stopped.
When you use Continue Response, the text is added to and merged into the last Assistant message. Sometimes, I'm unhappy with the content of the Continue Response, forcing me to edit the entire message, find the beginning of the new content, and remove it entirely, in order to retry. In effect, Open-WebUI lacks the ability to "retry a continued response" without losing the entire message.
Describe the solution you'd like
I would like an option to "Continue Response as a New Message", so that the continued text is distinctly separate from the prior, allowing it to be edited or retried without impacting the previous response that I'm happy with. I envision this as a User preference, changing the behavior of the existing Continue Response button rather than adding a new one.
Describe alternatives you've considered
Alternately, an User preference that unlocks/allows submitting a blank prompt to accomplish the same, essentially just asking the AI for more but keeping that new response separate from the last.
Additional context
"Just send a prompt to continue" might seem like a solution, but the AI response to pressing "Continue Response" and any sort of prompt to "Continue" or "Write more" can be very different.
Perhaps when back to back Assistent messages are passed to the model, they should be merged so the model believes it's one message. I am unsure if that matters.