[Feature Request]: Option to send generated image prompt instead of UI prompt to img2img

[space-nuko]

trafficstars

Is there an existing issue for this?

• [X] I have searched the existing issues and checked the recent builds/commits

What would your feature do ?

The "Send to img2img" button sends the prompt in the textbox to the img2img tab. I expect that what's sent to img2img is the prompt in the generated image. This is important for example if dynamic prompts is used and each image has a different prompt generated, I don't want to send the prompt template to img2img but the generated prompt instead, the randomization will interfere with the upscaling I want to do. Otherwise I have to keep remembering to go back to txt2img and copy the prompt every time

Proposed workflow

1. Generate txt2img
2. Select image in the gallery with prompt to use
3. Press "Send to img2img"
4. Prompt in generated infotext for the selected image is sent to img2img

Additional information

No response

[AlUlkesh]

That's not an issue with webui, but with the extension.

The way you describe it would be the default, but the extension changes it, so the wildcard prompt is transmitted: https://github.com/adieyal/sd-dynamic-prompts/blob/55d23cf9b3109a4685950486126de34631358a0c/sd_dynamic_prompts/callbacks.py#L53-L55

You might want to clear that up on the extensions's side. I don't know if this is a feature or a bug: https://github.com/adieyal/sd-dynamic-prompts

[space-nuko]

but the extension changes it, so the wildcard prompt is transmitted

That isn't correct, if I change the prompt textbox in txt2img and click Send, it changes the output in the img2img prompt. That happens when I disable dynamic prompts too. It doesn't seem to matter what image I select in the gallery, instead the Send functionality looks at the state of the UI (which would contain the raw prompt syntax). What I want is a toggle that uses the embedded PNG info for those gallery files instead of the UI for the state to send