Mass select and Delete/Save pics

Would be great to have an easier way to select and save pictures that you actually want. It is often the case that 90% of the pictures will be unused because of low resemblance or many other factors. So being able to select the ones you actually want and easily mass delete the others would be a nice UX improvement.

Filter Gallery by Model

I appreciate your work and your creativity but please can you work and finish the feature that you can filter the gallery by model. Its much more important that tiktok video features.

Mocap

Still can’t upload a source video. I can upload the photo just fine, but not the video.

not generating pictures

Several image generations appear as black placeholders since yesterday. The jobs completed but images never rendered. Could you please refund the credits for these failed generations

Upload audio files for talking videos

It would be great to be able to upload an audio file for a talking video, rather than choose one of the stock voices. The problem is that while there may be over 200 voices available, by the time you have filtered them for gender, language, age, etc. you often are left with only a few to choose from. Other AI services allow this feature, and it still should still be able to screen the files for appropriate content if necessary.

Forcing a reference photo for Hyper Realism

I’m curious how the model chooses which photo to reference strongest in created photos with Hyper Realism. There obviously is one, since things like clothes, hair, and location will resemble a particular uploaded photo if no prompt is given to change those attributes. It seems to be randomly chosen, as multiple models created with largely the same images will turn out wildly different, some much better than others based on the reference photo chosen. For example, if it decides to use primarily a close selfie photo, generated photos designed to be full body or on a longer focal lens will have low resemblance and look odd, and vice versa. And yet it will stick with that same reference photo even if you prompt with attributes clearly apparent in other uploaded photos. Confusingly, the reference photo will sometimes change randomly hours or days later, often to a worse photo, and no matter what you prompt you can’t get it to change back. Additionally, the primary reference photo is DIFFERENT if creating a photo with multiple models, and that reference photo is also persistent in all photos with multiple models. I’m sure you already know all this and specifically how it works, so I thought I’d ask to see if you know of a way to force the model to default to a particular reference photo chosen, either per model or per photo, instead of relying on random chance that it picks a decent one and that it won’t randomly change later. Either with a prompt (eg: “use photo in the forest as reference) or perhaps a certain image upload order, or ideally a “model edit” feature to allow for that customization. Thanks for reading!

New Edit Feature??

Could you add a feature for Nano Banana Pro that allows adding multiple reference images—essentially an editor mode? Right now, Magic Edit only applies to a single image

Zoom out feature not functioning correctly

Lately when I tried to use zoom out feature it simply takes the original photo and puts a weird frame around it (like it’s a movie poster or something) instead of actually zooming out to get a bigger perspective of the photo’s subject. Have lost a lot of credits this way.

Cannot login.

I have reactivate my account with email: andreas-patsim@hotmail.com. Then I paid yearly professional subscription but the app tells me to pay again

Magic edit

When I try to insert a different model into an existing photo by using the magic edit feature (ie I select the model I want to insert into the photo, then click magic edit and highlight the person In the photo that I want removed and replaced by the model I just selected for insertion) it doesn’t work. The resulting images do change the look of the person I attempt to replace, but it’s definitely not the model I selected to insert into image (it’s not any of my models for that matter).