I have been following this project for a while and love it. It is an excellent method of integrating the LLM into an interface which only has text-based inputs.
I recognize that each of the models are a bit different, however if it was possible, I make the recommendaiton that you integrate the OpenAI's /v1/models endpoint into your code, so that the user can interrogate the endpoint for available models. You might have to put an indicator next to models that your code confidently supports, and with the others being selectable, but with no assurance that it'll work.
I have been following this project for a while and love it. It is an excellent method of integrating the LLM into an interface which only has text-based inputs.
I recognize that each of the models are a bit different, however if it was possible, I make the recommendaiton that you integrate the OpenAI's /v1/models endpoint into your code, so that the user can interrogate the endpoint for available models. You might have to put an indicator next to models that your code confidently supports, and with the others being selectable, but with no assurance that it'll work.