The api can also modify images using prompts, can you add that to merlin itself as an option?

Gpt-Image-01 works great in bonkers, but it is full capabilities isn't use well if it's only usable as a image generator.

It's the most important part of gpt's-image-1, able to instruct and modify via text or instructed images

Like these for example, as directly advertised by openai:

With that, it can then be also used for mobile users, prompting is the greatest advantage to gpt image generation

Removing that function makes it look like a generic image generator that are only a bit better to other image generator