Feature Requests

Add Isometric and other types of camera projections to the Fiirefly camera list

Como profesor de idiomas genero imágenes como apoyo visual a pequeños diálogos o historias. El problema es que casi siempre necesito usar un mismo personaje en diferentes situaciones y al generar una imagen nueva el personaje cambia demasiado aunque lo describa bien. Sería muy útil poder generar un personaje con ciertas características y rasgos, y poder generar imágenes IA en donde ese personaje esté en diferentes situaciones, escenarios y acciones.

Provision to open images in new tab.

I have wanted to rotate the generated image to different angles before expanding in Firefly. I think especially in the cases of doing abstracted images, there are reasons to rotate to difgerent angle positions for aesthetic purposes. The generative expand would need to fill the non rectangular shaped rotated image.  At other times it is useful for rotating an image to a better horizontal position or to change the vertical of a point of interest. I'm sure this concept could be even more robust with other angle adjustment methods such as fixing keystoning and other kinds of distortion corrections. I want to expand after the rotation and before taking the image into PSD.

Unfortunately you are still way way way millions years way behind MidJourney .

When downloading the generated images, it would be nice if the style was included in the download name (e.g. graphic, hyperrealistic, Low Poly), etc.  

I tried the prompt "afo" which is common for people with weaker ankles. It's a lower leg brace. I also tried a brace prompt (and got casts). AFO is short for Ankle Foot Orthotics. I would love to make images for kids wearing afos, and adults to see their equipment normalized. 

There should be several categories to choose from. Each category needs its own algo. For example: there should be an algo for Coins then sub algos would be color. the algo should know when the prompt  is trying to make a realistic vs 3d coin. Also this Structure and Style uploads need to be more exacting to the prompts. 

Having a button that can convert an AI (Firefly) created image into a vector version would be helpful for a designer and save so much time - this is a better use than the Trace option, particularly in my field (for speed of output)

Please create a feature to download images as vectors for us graphic designers! 

i have an idea for structure reference, it will be more valuable if there is a draw tool on the structure reference section, simply a pencil to draw outline or any sketch directly on the go✌️

يرجى اضافة عنصر التشكيل لحروف اللغة العربية

It would be interesting to allow a rectangular selection tool to select a cropped portion of a generated image that might produce a great image on its own if it could simply be enlarged to a higher resolution than the crop by itself might yield. The crop would need to be regenerated in a way that keeps the same look and feel but produces higher resolution in a new version of the cropped image. In a way, it is creating the same image area in a new genertion with more detail and sharpness. Not sure if this makes sense.

It would make sense to include a good selection tool inside Firefly. The brush is not always effective in gathering what needs to be selected. A shape selection might be enough. However, it might be great to have additional tools for other types of selection.

日本語版を継続して使っている。プロンプトが反映されないことが多いので、反映されていない和文プロンプトをGeminiで英訳したところ反映された。このことはFireflyにおける和文読解力に問題があることの証左であり、日本語版についてはβ版レベルでしかないことの証明でもある。生成回数が無駄に増えるし英文の作成など手間がかかるので、日本において有償とするなら最低限改善すべきだ。

Please enable custom aspect ratios like in Midjourney. Thank you so much!

Vielen Dank. mfg

Uma ideia legal é colocar com que essa imagem já possa ir direto para o illustrator vetorizada, iria ser de grande ajuda 

Please make a hindi language command

He trabajado y probado esta fantástica aplicación desde hace ya muchos meses, veo con preocupación que han eliminado los parametros de fotografía como tipo de lente y apertura, hacen falta era uno de los valores importantes de este programa.

I have to create a series of images in one style of illustration. Once a style has been agreed upon.  There could be an opportunity to batch-submit prompts for the different content/subject matter to be executed in the exact same style. I'd say this is a common use of illustration or styles. You need to apply it to more than one deliverable or subject.I love it, though.

I read the info on the styles and structure sliders. It really doesn't feel very clear. Intensity and Strength? Perhaps some kind of visual representation would help clarify.

Can we paint over an object, pull a mask off of an object, or have the A.I. pick objects based on a grid picker of the image recognition, then create layer masks? Then assign GPUs to certain layers to render with different prompts. 

adjusting skin tone, or color tone on objects in generative would be useful, and rather than generating hoping for a result, having sliding bar to make live changes would be nice

I would really appreciate if there was an option to AI-convert a photo to vector illustration based on the user's photo and text prompt (and not just jpeg image compression areas)