[GH-ISSUE #874] [FEAT]: Persist LLM Providers parameters when toggling between providers #517

Closed
opened 2026-02-22 18:19:54 -05:00 by yindo · 4 comments
Owner

Originally created by @tiangao88 on GitHub (Mar 7, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/874

How are you running AnythingLLM?

Docker (remote machine)

What happened?

Running AnythingLLM on hosted version in Railway thanks to the deployment template (thanks!)

Parameters are not saved consistently which is quite an annoyance if you switch frequently LLM models to compare them
For example:

  • OpenAI : the API key is saved but the other parameters are not
  • LocalAI : the API key is saved but the other parameters are not
  • Perplexity: the API key is not saved but the other parameters are
  • Openrouter: none of the parameters are saved
  • Groq: perfect, all parameters are saved!

I haven't tried the other LLM providers.

Are there known steps to reproduce?

No response

Originally created by @tiangao88 on GitHub (Mar 7, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/874 ### How are you running AnythingLLM? Docker (remote machine) ### What happened? Running AnythingLLM on hosted version in Railway thanks to the deployment template (thanks!) Parameters are not saved consistently which is quite an annoyance if you switch frequently LLM models to compare them For example: - OpenAI : the API key is saved but the other parameters are not - LocalAI : the API key is saved but the other parameters are not - Perplexity: the API key is not saved but the other parameters are - Openrouter: none of the parameters are saved - Groq: perfect, all parameters are saved! I haven't tried the other LLM providers. ### Are there known steps to reproduce? _No response_
yindo added the bugDocker labels 2026-02-22 18:19:54 -05:00
yindo closed this issue 2026-02-22 18:19:54 -05:00
Author
Owner

@timothycarambat commented on GitHub (Mar 7, 2024):

The parameters are not saved at all? Or are not saved between restarts?

@timothycarambat commented on GitHub (Mar 7, 2024): The parameters are not saved at all? Or are not saved between restarts?
Author
Owner

@tiangao88 commented on GitHub (Mar 7, 2024):

The parameters are not saved at all? Or are not saved between restarts?

For example, I set up parameters for Openrouter and go test the LLM in a workspace.
I go back to the settings to switch to and set up OpenAI parameters and go try the LMM in a workspace.
I go back to the settings to switch back to Openrouter and all parameters are gone.

I just tried Azure OpenAI and all parameters are also not saved when switching back and forth different settings.

My Railway instance is set up to never go to sleep, and I never restart the instance.

@tiangao88 commented on GitHub (Mar 7, 2024): > The parameters are not saved at all? Or are not saved between restarts? For example, I set up parameters for Openrouter and go test the LLM in a workspace. I go back to the settings to switch to and set up OpenAI parameters and go try the LMM in a workspace. I go back to the settings to switch back to Openrouter and all parameters are gone. I just tried Azure OpenAI and all parameters are also not saved when switching back and forth different settings. My Railway instance is set up to never go to sleep, and I never restart the instance.
Author
Owner

@timothycarambat commented on GitHub (Mar 7, 2024):

Oh, this is actually normal. It ties back to a legacy reason as to why it happens but I agree that it is not an "ideal" behavior. I think we can make this change now given the current state of the codebase. So not a bug, but really a feature that would make toggling between providers not so much of a headache 👍

@timothycarambat commented on GitHub (Mar 7, 2024): Oh, this is actually normal. It ties back to a legacy reason as to why it happens but I agree that it is not an "ideal" behavior. I think we can make this change now given the current state of the codebase. So not a bug, but really a feature that would make toggling between providers not so much of a headache 👍
Author
Owner

@tiangao88 commented on GitHub (Mar 7, 2024):

Oh, this is actually normal. It ties back to a legacy reason as to why it happens but I agree that it is not an "ideal" behavior. I think we can make this change now given the current state of the codebase. So not a bug, but really a feature that would make toggling between providers not so much of a headache 👍

Ok, for some reason the parameters in the Groq LLM provider are saved and are sticky.
Therefore I was expecting that parameters in all other LLM providers would also be saved once and for all.

@tiangao88 commented on GitHub (Mar 7, 2024): > Oh, this is actually normal. It ties back to a legacy reason as to why it happens but I agree that it is not an "ideal" behavior. I think we can make this change now given the current state of the codebase. So not a bug, but really a feature that would make toggling between providers not so much of a headache 👍 Ok, for some reason the parameters in the Groq LLM provider are saved and are sticky. Therefore I was expecting that parameters in all other LLM providers would also be saved once and for all.
yindo changed title from [FEAT]: Persist LLM Providers parameters when toggling between providers to [GH-ISSUE #874] [FEAT]: Persist LLM Providers parameters when toggling between providers 2026-06-05 14:35:44 -04:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#517