Originally created by @Oaklight on GitHub (Mar 5, 2025).
Check Existing Issues
I have searched the existing issues and discussions.
Installation Method
Docker
Open WebUI Version
v0.5.19
Ollama Version (if applicable)
No response
Operating System
debian 12
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 checked the browser console logs.
I have checked the Docker container logs.
I have listed steps to reproduce the bug in detail.
Expected Behavior
When users have access to the Models page in their workspace, they should be able to customize/add their models and choose foundation models based on their preferences and availability.
This choice should be independent of the admin's model availability.
Actual Behavior
Models defined by user from workspace result in "404 model not found" if the chosen foundation model is not available for root admin. Users are configuring their own API. For example, deepseek-v3 as added in the following "logs and screenshots" that's only available to user API will trigger 404. But direct conversation with deepseek-v3 is working fine. Switching to yi-lightning that's available to both admin and user works fine for custom model.
Steps to Reproduce
admin has a few models from api
set to allow each user to use their own api
set to allow user to access workspace/models
user define their own custom model with choice of foundation models from their API model dropdown
run test conversation with such custom model
got 404
Logs & Screenshots
user defined model with deepseek-v3 only available to their api
user conversation with their model has 404
Additional Information
I use gateway to host lots of api providers, so on that api gateway, I see the query from openwebui has empty model name for such conversation test.
Originally created by @Oaklight on GitHub (Mar 5, 2025).
### Check Existing Issues
- [x] I have searched the existing issues and discussions.
### Installation Method
Docker
### Open WebUI Version
v0.5.19
### Ollama Version (if applicable)
_No response_
### Operating System
debian 12
### 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 checked the browser console logs.
- [x] I have checked the Docker container logs.
- [x] I have listed steps to reproduce the bug in detail.
### Expected Behavior
When users have access to the Models page in their workspace, they should be able to customize/add their models and choose foundation models based on their preferences and availability.
This choice should be independent of the admin's model availability.
### Actual Behavior
Models defined by user from workspace result in "404 model not found" if the chosen foundation model is not available for root admin. Users are configuring their own API. For example, deepseek-v3 as added in the following "logs and screenshots" that's only available to user API will trigger 404. But direct conversation with deepseek-v3 is working fine. Switching to yi-lightning that's available to both admin and user works fine for custom model.
### Steps to Reproduce
1. admin has a few models from api
2. set to allow each user to use their own api
3. set to allow user to access workspace/models
4. user define their own custom model with choice of foundation models from their API model dropdown
5. run test conversation with such custom model
6. got 404
### Logs & Screenshots
user defined model with deepseek-v3 only available to their api

user conversation with their model has 404

### Additional Information
I use gateway to host lots of api providers, so on that api gateway, I see the query from openwebui has empty model name for such conversation test.
GiteaMirror
added the bug label 2025-11-11 15:49:58 -06:00
If that's the design, then there should be no "+" option on workspace/models for user role. And the user who is granted access to the model defined by the admin should only be able to choose from a restricted selection of foundation models (for example let the admin curate a list of options) instead of seeing the user-side model list. or simply doesn't allow user role customization for workspace/models. Current UI design is confusing. @tjbck
@Oaklight commented on GitHub (Mar 5, 2025):
If that's the design, then there should be no "+" option on workspace/models for user role. And the user who is granted access to the model defined by the admin should only be able to choose from a restricted selection of foundation models (for example let the admin curate a list of options) instead of seeing the user-side model list. or simply doesn't allow user role customization for workspace/models. Current UI design is confusing. @tjbck
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @Oaklight on GitHub (Mar 5, 2025).
Check Existing Issues
Installation Method
Docker
Open WebUI Version
v0.5.19
Ollama Version (if applicable)
No response
Operating System
debian 12
Browser (if applicable)
No response
Confirmation
README.md.Expected Behavior
When users have access to the Models page in their workspace, they should be able to customize/add their models and choose foundation models based on their preferences and availability.
This choice should be independent of the admin's model availability.
Actual Behavior
Models defined by user from workspace result in "404 model not found" if the chosen foundation model is not available for root admin. Users are configuring their own API. For example, deepseek-v3 as added in the following "logs and screenshots" that's only available to user API will trigger 404. But direct conversation with deepseek-v3 is working fine. Switching to yi-lightning that's available to both admin and user works fine for custom model.
Steps to Reproduce
Logs & Screenshots
user defined model with deepseek-v3 only available to their api

user conversation with their model has 404

Additional Information
I use gateway to host lots of api providers, so on that api gateway, I see the query from openwebui has empty model name for such conversation test.
@tjbck commented on GitHub (Mar 5, 2025):
Custom models with direct connections are not supported.
@Oaklight commented on GitHub (Mar 5, 2025):
If that's the design, then there should be no "+" option on workspace/models for user role. And the user who is granted access to the model defined by the admin should only be able to choose from a restricted selection of foundation models (for example let the admin curate a list of options) instead of seeing the user-side model list. or simply doesn't allow user role customization for workspace/models. Current UI design is confusing. @tjbck