Description
While the AI is generating a response, the chat automatically scrolls down to follow the output. This makes it impossible to scroll up and read older messages at the same time — the view keeps jumping back to the bottom as new content is streamed in.
Steps to Reproduce
- Start a chat with a long message history
- Send a message and wait for the AI to start generating a response
- Try to scroll up to read earlier messages while the AI is still outputting
- The view keeps snapping back to the bottom, making scrolling upward impossible
Expected Behavior
- Auto-scroll and manual scrolling should be decoupled:
- If the user has not scrolled up: auto-scroll follows the AI output as usual
- If the user scrolls up during generation: auto-scroll is paused and the user can freely navigate the chat history
- Auto-scroll resumes automatically once the user scrolls back to the bottom (or after generation is complete)
This is a well-established pattern used by terminals, log viewers, and chat apps (e.g. Slack, Discord, VS Code terminal).
Actual Behavior
- Auto-scroll during AI output always overrides manual scrolling
- The user cannot read older messages while a response is being generated
- The only option is to wait until the AI finishes before scrolling
Why This Matters
- Long AI responses make this problem very noticeable
- Users often want to reference earlier context while the AI is still responding
- Forced auto-scroll feels uncontrollable and frustrating
Environment
- OpenGUI Desktop Interface
- OS: Linux
Description
While the AI is generating a response, the chat automatically scrolls down to follow the output. This makes it impossible to scroll up and read older messages at the same time — the view keeps jumping back to the bottom as new content is streamed in.
Steps to Reproduce
Expected Behavior
This is a well-established pattern used by terminals, log viewers, and chat apps (e.g. Slack, Discord, VS Code terminal).
Actual Behavior
Why This Matters
Environment