Custom LLM needs OAuth setup per assistant
# support
p
Currently I can only add Custom LLM OAuth URL and Credentials in Provider Keys, but every time I want to test different environment (local, staging, prod) I have to switch URL instead of having assistants for each environments. Cant even update URL, getting error, have to delete credentials and add again.
v
I understand your frustration with having to delete and re-enter OAuth credentials when switching between environments for Custom LLMs in Vapi. According to the documentation, OAuth for Custom LLMs is currently configured through Provider Keys with a single URL configuration. While I don't see a built-in way to have environment-specific assistants, I'd like to better understand your workflow. Have you tried creating separate credential entries with different names (like "Custom LLM - Dev", "Custom LLM - Staging") to manage your environments, or is there a limitation preventing this approach? Sources: - [OAuth 2 Authentication for Custom LLMs and Webhooks](https://docs.vapi.ai/changelog/2024/12/6) - [Submitting Feature Requests](https://docs.vapi.ai/support)
p
Did it through API, but having all settings in UI would be nice.
s
Please provide details of your issue and steps taken to resolve it. This will help me understand your challenges and tailor solution and then pass it to the team.
p
1. Go to assistant UI 2. No settings in assistant to add authentication credentials 3. Go to https://github.com/VapiAI/server-sdk-python 4. Link to API reference not working 5. Look around in docs 6. Write script to update assistant settings 7. Updated settings Expected behavior: 1. Go to assistant UI 2. Update settings
s
if you would like to request that feature, please add it here as I'm sure many would benefit from this addition: API is the best way to build as of right now. you get full access while we build out the UI in the Dashboard. I was looking at some other threads and some others have mentioned multiple environment management as well so I suggest looking for the feature request and upvoting it or if you can't find it, creating one and detailing what you would like to see.