[GH-ISSUE #5291] [BUG]: ALLM window cuts off text instead of wrapping it. #5000

Closed
opened 2026-06-05 14:51:24 -04:00 by yindo · 7 comments
Owner

Originally created by @bulrush15 on GitHub (Mar 28, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5291

Originally assigned to: @timothycarambat on GitHub.

How are you running AnythingLLM?

Desktop App

What happened?

I have upgraded to ALLM v1.11.2. I tend to make my windows smaller on my desktop in Windows 11. I'm using a Qwen3:8b model. My Nvidia GPU has 12GB of VRAM.

When I make the window small enough to see past chats in a workspace the text is cut off, it does not wrap. Both my question and AI answers should wrap to the window size.

Screen shot:

Are there known steps to reproduce?

  1. Start ALLM.
  2. Open a workspace with existing questions and answers that uses the Qwen3:8b model.
  3. Make the window skinnier. The text will not wrap when the ALLM main window gets too skinny.
Originally created by @bulrush15 on GitHub (Mar 28, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5291 Originally assigned to: @timothycarambat on GitHub. ### How are you running AnythingLLM? Desktop App ### What happened? I have upgraded to ALLM v1.11.2. I tend to make my windows smaller on my desktop in Windows 11. *I'm using a Qwen3:8b model.* My Nvidia GPU has 12GB of VRAM. When I make the window small enough to see past chats in a workspace the text is cut off, it does not wrap. Both my question and AI answers should wrap to the window size. Screen shot: <img src="https://github.com/user-attachments/assets/7a0c867b-db2a-4db7-818f-b3caae08fe81" /> ### Are there known steps to reproduce? 1. Start ALLM. 2. Open a workspace with existing questions and answers that uses the Qwen3:8b model. 3. Make the window skinnier. The text will not wrap when the ALLM main window gets too skinny.
yindo added the possible bugneeds info / can't replicateDesktopinvestigating labels 2026-06-05 14:51:24 -04:00
yindo closed this issue 2026-06-05 14:51:24 -04:00
Author
Owner

@timothycarambat commented on GitHub (Mar 28, 2026):

Compressed Sidebar and small window - not seeing this. Might be a really long link - which we do not break-word on.

Image

Will have to investigate if this is a windows thing or not, but the windows should render the same

<!-- gh-comment-id:4149208506 --> @timothycarambat commented on GitHub (Mar 28, 2026): Compressed Sidebar and small window - not seeing this. Might be a really long link - which we do not `break-word` on. <img width="551" height="1106" alt="Image" src="https://github.com/user-attachments/assets/5c4a1fa4-9dba-4f75-ac61-e25fc4ae2019" /> Will have to investigate if this is a windows thing or not, but the windows should render the same
Author
Owner

@bulrush15 commented on GitHub (Mar 29, 2026):

This only happens when I'm using a Qwen model. It does not happen when using llama3.2 or in my RAG AI using llama3.2:3b. Try using a Qwen3:8b model.

Could this issue be model-related? Will a specific model return CSS-styled text that has a fixed length line and break point?

<!-- gh-comment-id:4149989485 --> @bulrush15 commented on GitHub (Mar 29, 2026): This only happens when I'm using a Qwen model. It does not happen when using llama3.2 or in my RAG AI using llama3.2:3b. Try using a Qwen3:8b model. Could this issue be model-related? Will a specific model return CSS-styled text that has a fixed length line and break point?
Author
Owner

@timothycarambat commented on GitHub (Mar 30, 2026):

@bulrush15 No, definitely not model specific. Sure some markdown could have an output style that is causing that but its not a model specific thing

<!-- gh-comment-id:4158336080 --> @timothycarambat commented on GitHub (Mar 30, 2026): @bulrush15 No, definitely not model specific. Sure some markdown could have an output style that is causing that but its not a model specific thing
Author
Owner

@timothycarambat commented on GitHub (Apr 6, 2026):

I have tried over and over to get this to happen on my windows machine. Can you share a specific prompt output (pencil icon lets you see raw text) so I can try to see if this is some encoding or text width thing?

<!-- gh-comment-id:4193832711 --> @timothycarambat commented on GitHub (Apr 6, 2026): I have tried **over and over** to get this to happen on my windows machine. Can you share a specific prompt output (pencil icon lets you see raw text) so I can try to see if this is some encoding or text width thing?
Author
Owner

@bulrush15 commented on GitHub (Apr 6, 2026):

  1. Remember to use a narrow window for ALLM.
  2. Model: Qwen3:8b
  3. Here's close to the prompt I used above: What is the market share of the programming language Python?
    I have confirmed that when I make the ALLM more narrow the answer text does not wrap.

Wider window wraps properly.

Image

Narrow window does not wrap response text.

Image

In some web pages if text is put into a fixed width table with a minimum width I see the same problems in a narrow windw.

<!-- gh-comment-id:4193959282 --> @bulrush15 commented on GitHub (Apr 6, 2026): 1. Remember to use a narrow window for ALLM. 2. Model: Qwen3:8b 3. Here's close to the prompt I used above: What is the market share of the programming language Python? I have confirmed that when I make the ALLM more narrow the answer text does not wrap. Wider window wraps properly. <img width="475" height="349" alt="Image" src="https://github.com/user-attachments/assets/d6f08f71-4b8e-45a4-9533-c3169db5b336" /> Narrow window does not wrap response text. <img width="638" height="388" alt="Image" src="https://github.com/user-attachments/assets/c6c9dd84-d1a0-4753-abda-5e7f1c1b2fe3" /> In some web pages if text is put into a fixed width table with a minimum width I see the same problems in a narrow windw.
Author
Owner

@bulrush15 commented on GitHub (Apr 24, 2026):

I upgraded to ALLM v1.12.1 and my workspaces are still there.

This was not fixed in ALLM 1.12.1. The problem is not just with a long URl but with long text with words separated by spaces.

  1. Should I open another issue for this?
  2. Do I have to delete the workspace that uses Qwen3 and remake it?

I use LLMs of Llama 3.2 and Qwen3:8b and it only seems to be a problem in Qwen3:8b.

Is there another issue for this? Can someone add me to the other issue so I get updates when they are made to the Github issue?

<!-- gh-comment-id:4313679169 --> @bulrush15 commented on GitHub (Apr 24, 2026): I upgraded to ALLM v1.12.1 and my workspaces are still there. This was not fixed in ALLM 1.12.1. The problem is not just with a long URl but with long text with words separated by spaces. 1. Should I open another issue for this? 2. Do I have to delete the workspace that uses Qwen3 and remake it? I use LLMs of Llama 3.2 and Qwen3:8b and it only seems to be a problem in Qwen3:8b. Is there another issue for this? Can someone add me to the other issue so I get updates when they are made to the Github issue?
Author
Owner

@bulrush15 commented on GitHub (Apr 24, 2026):

I have made a new workspace in ALLM v1.12.1 using Qwen3:8b and I still see this problem of cutting off text.

Icons in the area where I enter a prompt are also cut off in a narrow ALLM window.

Image
<!-- gh-comment-id:4313731972 --> @bulrush15 commented on GitHub (Apr 24, 2026): I have made a new workspace in ALLM v1.12.1 using Qwen3:8b and I still see this problem of cutting off text. Icons in the area where I enter a prompt are also cut off in a narrow ALLM window. <img width="353" height="806" alt="Image" src="https://github.com/user-attachments/assets/4c9dd988-9c29-44be-8ef6-29e410b17bd1" />
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#5000