Is your feature request related to a problem? Please describe.
Now you have a dropdown of models, which you have to register one by one.
Most inference providers have many models, and those come and go every week.
When I generate an API token for an inference provider, it is visible only once. Therefore, to register a new model from the same provider into OdooTerminal, I need to generate a new token.
Describe the solution you'd like
Instead of defining models one by one, just define inference provider connections. Providers have an endpoint that lets us know available models, so that's where the list of models should come from.
For providers with thousands of models, it would help a lot if we can have a quick search, favorite models, and even be able to pick a default one (or just continue always using the last one used if still available).
Describe alternatives you've considered
Store the API token elsewhere so I don't need to create a new one each day, and keep the model list updated by hand.
Additional context
One image speaks 1000 words 😆
Is your feature request related to a problem? Please describe.
Now you have a dropdown of models, which you have to register one by one.
Most inference providers have many models, and those come and go every week.
When I generate an API token for an inference provider, it is visible only once. Therefore, to register a new model from the same provider into OdooTerminal, I need to generate a new token.
Describe the solution you'd like
Instead of defining models one by one, just define inference provider connections. Providers have an endpoint that lets us know available models, so that's where the list of models should come from.
For providers with thousands of models, it would help a lot if we can have a quick search, favorite models, and even be able to pick a default one (or just continue always using the last one used if still available).
Describe alternatives you've considered
Store the API token elsewhere so I don't need to create a new one each day, and keep the model list updated by hand.
Additional context
One image speaks 1000 words 😆