Feature Requests

I think a feature that allows retaining the same uniform clothing while switching characters is needed. This would enable me to create designs for characters in the same series without having to repeatedly test and adjust during the process.

Can you add a function in the new editor to delete the pixels completely in the masked area and have the AI rebuild them, like in Canvas? This method works better to polish images.

An editor that lets us start with a MALE or FEMALE generic voice then EDIT it by choosing "deeper or higher" with a toggle 1-10 level to reach desired tone, twangy level, british-american, etc. each with 1-10 ONCE VOICE IS CREATED it can be saved similar to character mode where it can then be used WITH TEXT to create a talking photo. ESPECIALLY if were able to load a sample clip for it to emulate similar to "image reference , image - image" MAYBE voice reference VOICE - VOICE Maybe not even a video, just a talking GIF would be great if were able to create the voice.

Since you've added flux dev I haven't looked back, but only being able to output at 768x960 is making me wish we had even just a 2x boost on the dimensions without extreme quality loss of a 4x 8x.

Let us decide if we want a trained model to be active by including its name in the prompt or something

Allow a prompt input so that we can direct what we want the AI to vary.

Since there are various angle prompts "side view" "profile view" ETC... and not sure which are best for getimg how about below the PROMPT FIELD where you have RANDOM PROMPT it would be great to have a pull down button where we could choose the following... SIDE VIEW - SNAIL VIEW - BIRD VIEW - OVERHEAD - REAR VIEW - OVER SHOULDER - UNDERNEATH

For those of us who use MODEL TRAINER alot, my consistent characters have 100 and it would be great to locate the name in alphabetical order. Maybe a toggle button between NEW and ALPHABET