Feature Requests

Firefly is great at expanding images, but if you start with a horizon that is slightly crooked, the results are often not good. Asking the AI to straighten the horizon often leaves an offset. It would be great to have an option to straighten the image manually

As mention please add Please add in Seedance 2.0 and add references with long prompts abilities

La idea es simple, que agreguen motores de render como Seedance 2.0 en adobe firefly, asi como ya esta Kling 3.0 o omni, me gustaria que agreguen Seedance 2.0. Tambien podrian agregar google omni

Firefly text to speech needs a special setting for slow, well-articulated speech for English learners, and the current interface does not allow for such very precise and careful enunciation. So when adjusting for tone, why not have such a setting so that individual words are emphasized that are particularly helpful for those learning to pronounce properly, for educational or therapeutic purposes. Thank you.

When upscaling an image with Firefly, we should be able to compare the results before and after upscale.

I would like to request improvements to the video generation workflow in Boards, especially for the Seedance models.Currently, when generating videos with the Seedance model inside Boards, there appears to be only a first-frame option available. There is no clear option to upload or use additional media references, such as source images, video clips, or other visual assets, during the generation process.It would be very helpful to add more media upload options for Seedance video generation in Boards, such as:Uploading reference images Uploading video clips Adding start and end frames Using multiple visual references Supporting image-to-video and video-to-video workflowsI would also like to request more aspect ratio options for the ChatGPT image/video model and Seedance models. Additional presets such as 1:1, 4:5, 3:4, 4:3, 16:9, 9:16, 2:3, 3:2, and 21:9, along with custom aspect ratio support, would make the tools much more flexible for social media, presentations, cinematic videos, an

Add a second prompt box that can be toggled on and off. The second prompt box is a Style Reference so you dont need to include the style text in the prompt and you dont use character space in the prompt box, allowing  for better prompts since style text can be added in a second box.

gostaria de saber se é possível adicionar ao aprimoramento de imagem o modelo Magnific V2?Tem uma excelente qualidade de detalhe e resolução das texturas.

The image generation really needs multiple chats so you can keep everything in order. Making everything in one causes problems with organization, and my browser crashes occasionally with not enough memory.When creating images we need to have different active chats so organizing what we are making is more feasible, and creates a lot less clutter. Also it will run smoother.

Zur Zeit erlaubt Seedance nicht die Verwendung menschlicher Charaktere. Das macht die Funktion für mich völlig unbrauchbar, da ich Sequenzen für Musikvideos generiere. Was nutzt mir Charakterkonsistenz, wenn ich die Charaktere nicht definieren kann? Meine Charakterbilder sind ebenfalls KI generiert.

There should be multiple generations available for higher tiers. I have the premium and I can generate only one photo or video at the time, which makes it very difficult to produce anything in a reasonable time. 

I generate lots of videos with audio but I don’t want music in them. If I forget to add that in the prompt, I need to regenerate the videos. I suggest the audio toggle to be split into Sound FX and music. 

Google just released Nano Banana 2 Lite a couple of days ago, and I've already seen it integrated into several AI image generation platforms. Does anyone know when Nano Banana 2 Lite will be available in Adobe Firefly?Thanks!

es crear imagenes con efecto de semitono con imagenes  vectorizadas. para diferentes fondos 

The text field for writing a prompt is too short. It doesn't have enough space to write a detailed prompt. Please make the text field larger

Stickers or pins for marking selected content to find it later on a board. Each sticker can have tags field to make the search easierSome utility instruments for sorting content

Para tener un mejor control del gasto de tokens, sugiero implementar las siguientes mejoras: Visualización del total: Junto al coste de tokens de cada generación, podría mostrarse también el consumo total acumulado (o el saldo restante). Cobro por descarga: Como ya mencionó otro usuario, apoyo que sería ideal que los tokens solo se consumieran al descargar la imagen o el vídeo terminado. Coste dividido (Alternativa): Si lo anterior no fuera posible, una buena opción sería que la generación inicial consuma solo una parte de los tokens, y el resto se cobre únicamente si el usuario decide descargar el recurso. Muchas gracias!😁

Siendo miembro de Adobe, ya con una suscripción mensual al pack adobe pro, sería muy útil la posibilidad de compra de créditos extra para generación en firefly sin la necesidad de contratar otra suscripción mensual, sino comprar por paquete de crédito, y sin restricción de tiempo de uso de dicho crédito. ¡Gracias!

Using the batch image processing. Results are great however I am processing 90+ headshots at a time. File names are the persons name, but the output files are changed to a long string of letters and numbers. I now have no idea who is who and manually renaming them by cross referencing the original files is a loss of all time saved. Plus the human error risk is high in this kind of work. Please Allow AI assistant batch processing to preserve the original source filename on output files.

生成履歴の画面で、チェックボックスを付けるとお気に入りの星マークが消えてしまうのですが、お気に入り以外を削除する作業がしづらいので、以前のように戻してほしい。以前はチェックを付けても、お気に入りの星マークが消える事はありませんでした。[Moved to Adobe Firefly Feature Requests by a moderator]

Considero que las imágenes deben cobrarse o gastar créditos una vez descargadas para su uso final,  porque aunque la imagen se genere no necesariamente se usa; esto debido a que los clientes solicitan cambios de imágenes con frecuencia durante el proceso de diseño. Debería implementarse un sistema de marcas de agua que permitan probar las fotos antes de descargarlas, (como se hace en las fotos de stock) de esta manera se pueden probarse varias.

It is unfair that I am charged for ai misunderstanding a prompt, this would be solved by only being charged for videos that are downloaded

Previously we were able to delete multiple files at once:Go to the 25 or 30second mark.Now we can only delete one file at a time.Could we go back to being able to delete multiple files at once?Short cutchoose one image, then the next row down or up select an image.hold shift + clickThank youNate

Bonjour, Je trouve vraiment dommage la suppression du modèle d'image de luciole 2 / Firefly 2 depuis la dernière mise à jour. Le rendu des images était  bien spécifique au niveau du style, des textures, de la douceur des lumières. Le côté dessiné à la main ne se retrouve ni avec Firefly 3, ni avec Firefly 4, ni avec Firefly 4 ultra. Le résultat est totalement différent.  J'avais commencé un projet avec Firefly 2  et il collait parfaitement à ce que je recherchais en qualité d'image. J'aurais trouvé judicieux de le laisser pour avoir un plus large panel de styles visuels.  Mon projet tombe à l'eau à cause de ce changement. Pensez-vous à l'éventualité de le réintégrer ? Merci d'avance pour votre réponse. Bien à vous,Nora C

As Adobe Boards evolve into large-scale infinite canvases for creative collaboration, we often struggle to relocate explorations, assets, feedback, or earlier iterations across sprawling boards. Introducing a persistent “Board Map Navigator” icon similar to Miro’s minimap experience would provide a real-time bird’s-eye view of the entire canvas, enabling users to instantly orient themselves, jump to specific sections, search and highlight content, track collaborators, and bookmark key exploration areas. This would dramatically improve discoverability, reduce navigation fatigue, streamline review workflows, and make large AI-driven creative boards significantly more intuitive and manageable for teams working across complex visual storytelling and ideation environments.