Feature Request: OpenAI/OpenAI compatible API integration

[Todokete]

The ability to supply an API key and select a GPT, Claude, or Gemini model to prompt would be much appreciated

[StableLlama]

I've just used Gemini to caption a batch of images with a ComfyUI workflow - and I must say it was the best and easiest captioning experience I had. It was basically only installing of https://github.com/AINxtGen/ComfyUI-GeminiAPI and then go.

So yes, integrating it in taggui would be highly appreciated!