issue: Long model names and paths are unreadable #5547

Closed
opened 2025-11-11 16:24:01 -06:00 by GiteaMirror · 7 comments
Owner

Originally created by @mrkrsl on GitHub (Jun 15, 2025).

Check Existing Issues

  • I have searched the existing issues and discussions.
  • I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

0.6.14

Ollama Version (if applicable)

No response

Operating System

All (desktop + mobile#

Browser (if applicable)

No response

Confirmation

  • I have read and followed all instructions in README.md.
  • I am using the latest version of both Open WebUI and Ollama.
  • I have included the browser console logs.
  • I have included the Docker container logs.
  • I have provided every relevant configuration, setting, and environment variable used in my setup.
  • I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
  • I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
  • Start with the initial platform/version/OS and dependencies used,
  • Specify exact install/launch/configure commands,
  • List URLs visited, user input (incl. example values/emails/passwords if needed),
  • Describe all options and toggles enabled or changed,
  • Include any files or environmental changes,
  • Identify the expected and actual result at each stage,
  • Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

The model name in the UI (in all locations) should be meaningful and readable when connecting Open WebUI to an external API.

All locations means: in the model selector widget (top left), on the new chat page, on the chat page as the label of the model above its replies.

There is no place in the Open WebUI Settings to edit or otherwise control this string, which is rendered directly from the API.

Actual Behavior

When connected to my local API, the model names are extremely long and include the file path, meaning they don't fit in the UI and are unreadable. This means they are not meaningful strings and look cluttered. This is as true on desktop, where there's a lot of horizontal space as it is on mobile.

All I see for all model replies is the beginning of the file path to the model name, as delivered by the server (which I cannot change). You'll notice it only displays '/Users/mark/.cache/sanctu' where it is cut off.

See screenshot.

Steps to Reproduce

  1. Connect a local API in connections. (Mine is served by an app called Sanctuary, which runs equivalently sized models significantly faster than the docker ollama integration.)
  2. Load the chat UI.
  3. Note the extremely long model name, including the file path, wherever it occurs in the UI.

Note that I have no control of the model name sent by the Sanctuary API.

See screenshot attached.

Logs & Screenshots

See screenshots. Note model name top left in the UI and also above the input box and above every message. (Underlined in yellow.)

Image

Image

Additional Information

No response

Originally created by @mrkrsl on GitHub (Jun 15, 2025). ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version 0.6.14 ### Ollama Version (if applicable) _No response_ ### Operating System All (desktop + mobile# ### Browser (if applicable) _No response_ ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior The model name in the UI (in all locations) should be meaningful and readable when connecting Open WebUI to an external API. All locations means: in the model selector widget (top left), on the new chat page, on the chat page as the label of the model above its replies. There is no place in the Open WebUI Settings to edit or otherwise control this string, which is rendered directly from the API. ### Actual Behavior When connected to my local API, the model names are extremely long and include the file path, meaning they don't fit in the UI and are unreadable. This means they are not meaningful strings and look cluttered. This is as true on desktop, where there's a lot of horizontal space as it is on mobile. All I see for all model replies is the beginning of the file path to the model name, as delivered by the server (which I cannot change). You'll notice it only displays '/Users/mark/.cache/sanctu' where it is cut off. See screenshot. ### Steps to Reproduce 1. Connect a local API in connections. (Mine is served by an app called Sanctuary, which runs equivalently sized models significantly faster than the docker ollama integration.) 2. Load the chat UI. 3. Note the extremely long model name, including the file path, wherever it occurs in the UI. Note that I have no control of the model name sent by the Sanctuary API. See screenshot attached. ### Logs & Screenshots See screenshots. Note model name top left in the UI and also above the input box and above every message. (Underlined in yellow.) ![Image](https://github.com/user-attachments/assets/c7b2dd08-4d79-42bb-bd28-cc5b684f8360) ![Image](https://github.com/user-attachments/assets/0fe6df1d-de93-4439-8492-15e7b2dda45e) ### Additional Information _No response_
GiteaMirror added the bug label 2025-11-11 16:24:01 -06:00
Author
Owner

@mrkrsl commented on GitHub (Jun 15, 2025):

Here's a desktop screenshot that shows how silly the long model names look:

Image
@mrkrsl commented on GitHub (Jun 15, 2025): Here's a desktop screenshot that shows how silly the long model names look: <img width="1031" alt="Image" src="https://github.com/user-attachments/assets/60eb3d7e-c27c-4e34-ab4e-18f2aefc33ae" />
Author
Owner

@tjbck commented on GitHub (Jun 16, 2025):

Image

Full model name should be displayed on hover with dc3055873bba6e36e88cba304ad9f3cd89de50ad!

@tjbck commented on GitHub (Jun 16, 2025): ![Image](https://github.com/user-attachments/assets/f5bea20f-144f-4768-8d71-45be84939742) Full model name should be displayed on hover with dc3055873bba6e36e88cba304ad9f3cd89de50ad!
Author
Owner

@mrkrsl commented on GitHub (Jun 16, 2025):

That only partially solves this issue. The long string, when there is one, still litters the UI and looks bad. That's especially true in a chat since the model name is written out with every reply from the model.

There should be a way of renaming, or otherwise controlling the displayed version of the model name.

@mrkrsl commented on GitHub (Jun 16, 2025): That only partially solves this issue. The long string, when there is one, still litters the UI and looks bad. That's especially true in a chat since the model name is written out with every reply from the model. There should be a way of renaming, or otherwise controlling the displayed version of the model name.
Author
Owner

@Classic298 commented on GitHub (Jun 16, 2025):

@mrkrsl you can rename models in the Model settings in admin panel?

@Classic298 commented on GitHub (Jun 16, 2025): @mrkrsl you can rename models in the Model settings in admin panel?
Author
Owner

@mrkrsl commented on GitHub (Jun 16, 2025):

@mrkrsl you can rename models in the Model settings in admin panel?

Ah! Indeed you can. Thank you.

However, it makes me wonder why these core admin screens are so deeply hidden. It's the first time I've found them. (And believe me, I hunted for them!) Surely these core settings should be more accessible?

@mrkrsl commented on GitHub (Jun 16, 2025): > [@mrkrsl](https://github.com/mrkrsl) you can rename models in the Model settings in admin panel? Ah! Indeed you can. Thank you. However, it makes me wonder why these core admin screens are so deeply hidden. It's the first time I've found them. (And believe me, I hunted for them!) Surely these core settings should be more accessible?
Author
Owner

@tjbck commented on GitHub (Jun 16, 2025):

@ayanahye could you write a tutorial for common often used tasks these?

@tjbck commented on GitHub (Jun 16, 2025): @ayanahye could you write a tutorial for common often used tasks these?
Author
Owner

@mrkrsl commented on GitHub (Jun 16, 2025):

Final thoughts, in case they help: Your Admin/User model makes sense, and I see that these settings fall on the Admin side. But it was totally non-obvious that the Admin settings are there, at least to me. I had visited the top-level of the Admin panel after first installing, but I totally missed the menu at the top (which is light grey against white, almost like those options are disabled.) It remained non-obvious after the above comment led me to them, and I still had to hunt. So maybe it's a matter of making these much more obvious.

I'd been looking for other things too, which I also just found in Admin Panel > Settings, like turning off the prompt hints. So, my mistake for not finding these key admin panels, but it feels like they've been hidden.

Hope this helps.

@mrkrsl commented on GitHub (Jun 16, 2025): Final thoughts, in case they help: Your Admin/User model makes sense, and I see that these settings fall on the Admin side. But it was totally non-obvious that the Admin settings are there, at least to me. I had visited the top-level of the Admin panel after first installing, but I totally missed the menu at the top (which is light grey against white, almost like those options are disabled.) It remained non-obvious after the above comment led me to them, and I still had to hunt. So maybe it's a matter of making these much more obvious. I'd been looking for other things too, which I also just found in Admin Panel > Settings, like turning off the prompt hints. So, my mistake for not finding these key admin panels, but it feels like they've been hidden. Hope this helps.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#5547