Support direct image editing endpoints like https://api.openai.com/v1/images/edits

What's happening?

Currently Big-AGI seems to support image only models through /draw command. If a model like gpt-image-2.5 supports editing, currently you can access it indirectly through chat model like Astra, while enabling "Images" option.

This seems to go through Response API and Astra ends up rewriting the prompt that goes into underlying edit request to the actual image model.

To have direct control over editing prompt, a request needs to be sent to https://api.openai.com/v1/images/edits endpoint.

For documentation you can see https://developers.openai.com/api/reference/resources/images/methods/edit

I propose adding /edit command to the chat that would edit last image with the specified prompt.

OpenRouter presumably supports image editing by providing reference images to input_references parameter to regular /images endpoint: https://openrouter.ai/docs/api/api-reference/images/generate-an-image#body-input-references.

(FWIW, this is low priority issue for me, as I typically generate images via my own scripts to get multiple variations of the image quickly)

Where does this happen?

Big-AGI Pro (big-agi.com)

Impact on your workflow

Low - Minor inconvenience

Environment (if applicable)

No response

Additional context

No response