[GH-ISSUE #7283] Autosave response nodes every 10 seconds during text generation, to guard against data loss from crashes. #14687

Closed
opened 2026-04-19 20:59:27 -05:00 by GiteaMirror · 0 comments
Owner

Originally created by @barryd15 on GitHub (Nov 23, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/7283

Feature Request

Is your feature request related to a problem? Please describe.
When I experience crashes of various kinds, including browser crashes, python server crashes and complete PC crashes, I lose all text already generated from a running response generation.
Due to issues with my PC & OS lately, crashes have been quite frequent, and it hurts every time I lose a high-quality generation near the end, due to a crash. It's particularly painful when the generations take a long time, for example 30 minutes.

Describe the solution you'd like
During response generation, the new response node and the streamed text within should be periodically saved. A default of 10 seconds worked fine for me in a similar feature on another LLM UI.
This feature could be optional and the period configurable.
But consider switching it on by default, too. It could save someone without them even knowing the feature exists.

Describe alternatives you've considered
I plan to address this by fixing my PC, but that's an ongoing project. I think it's a good feature idea that will help others.

Additional context
Add any other context or screenshots about the feature request here.

Originally created by @barryd15 on GitHub (Nov 23, 2024). Original GitHub issue: https://github.com/open-webui/open-webui/issues/7283 # Feature Request **Is your feature request related to a problem? Please describe.** When I experience crashes of various kinds, including browser crashes, python server crashes and complete PC crashes, I lose all text already generated from a running response generation. Due to issues with my PC & OS lately, crashes have been quite frequent, and it hurts every time I lose a high-quality generation near the end, due to a crash. It's particularly painful when the generations take a long time, for example 30 minutes. **Describe the solution you'd like** During response generation, the new response node and the streamed text within should be periodically saved. A default of 10 seconds worked fine for me in a similar feature on another LLM UI. This feature could be optional and the period configurable. But consider switching it on by default, too. It could save someone without them even knowing the feature exists. **Describe alternatives you've considered** I plan to address this by fixing my PC, but that's an ongoing project. I think it's a good feature idea that will help others. **Additional context** Add any other context or screenshots about the feature request here.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#14687