Feature Requests

Hi, last week, I only could use one reference image in Gemini 2.5 Nanobananana (now vn 3 also).This week, it seems that I can use more than one reference image; and that is great for some tasks but it increases my error rate for my current task:I have to remember to delete the orginal reference image and apply a second one;otherwise, there is the original redundant reference image beside the intended reference image.Could you have a mode to switch from single image reference to multi please?  Thanks!

When I generate a video, I usually start with a quick test using a fast model to fine-tune the prompt before switching to a pro model. For example, I may want a 4-second video and run the tests with Veo 3.1 Fast, but when I switch to the pro model, the duration automatically changes to 8 seconds. This has happened to me several times: I forget to adjust the duration and end up consuming twice the credits.Related to this, it has also happened to me on multiple occasions that, while waiting for the video to be generated, during that time for reflection, I realize that some setting is not quite right or that I have a new idea. During generation, it is indicated that if I close the tab, the generation will be interrupted and I will lose the video, but it is not clear whether the credits are also consumed in that case.If the credits are indeed consumed, it might be useful to have a button to cancel the video generation, consuming only half of the credits.

I’ve noticed that the speaking quality of Text-to-Voice-over is better than that of Text-to-Avatar. The pronunciation in Dutch sounds clearer, more natural, and aligns better with how the language is supposed to be spoken. Could this be because Text-to-Voice-over uses ElevenLabs, while Text-to-Avatar does not?If so, is it possible for Text-to-Avatar to also be powered by ElevenLabs?

I want the videos to be much longer at least 10 to 20 minutes

Increase the characters / word input for the prompt as there are huge prompt to generate specific images

Hallo Adobe-Team,ich hätte einen Vorschlag zur Verbesserungder Textintegration in Firefly:Derzeit generiert die KI Text in Bildern oft automatisch, was zu verzerrten oder unleserlichen Schriftzügen führen kann.Mein Vorschlag:• Ermöglicht es dem Nutzer, konkrete Textarten / Fonts / Layouts auszuwahlen, bevor der Text in das Bild eingefügt wird.•Die Kl passt den Text anschließend automatisch perspektivisch, farblich und proportional an das Bild an.Nutzen:Höhere Lesbarkeit und sauberes DesignMehr Kontrolle für kreative NutzerBesonders nützlich für Poster, Logos, Social Media Conte-oder Marketingmaterial

you need a deeper bassier voice for the txt to speach

I'd like to provide feedback on the user experience when accessing my image history in Firefly. The current navigation flow is causing significant frustration and disrupting my workflow.Specific Issues:Confusing navigation between Canvas and History The path from the main canvas to viewing my image history feels unintuitive and awkward. It's not clear or easy to access my past work.Viewing images requires too many steps When I click on an image from my history to verify it's the one I want, getting back to my history gallery requires unnecessary extra clicks. This back-and-forth should be seamless but instead feels clunky."Back to Gallery" button goes to the WRONG gallery This is the most frustrating issue: When I'm viewing one of MY images and want to return to MY gallery, the first/default option takes me to the PUBLIC gallery instead—where none of my images are stored. Why would the navigation default to a public gallery when I'm working with my personal images?Suggested Improvement

A opção português não aparece em escolher idioma, para narração 

To generate the image to video, when i add my image in I want the characters to dance. the AI does a good job doing that. but then I have to find a song that might work with how they are dancing. we should be able to add the name, or link to the song in each variation we create so they can move or dance to the song correctly. that would be fun.

It'd be great to have more control over objects and camera angles. If I can select an object, then use the familiar controls from After Effects such as scale, rotate, and position to move objects the way I want that would be amazing. Same thing for camera control, I'd like to direct the camera to look at things when I want them to. Or if I can move an image in 3D space and have the AI interpret it as a person walking or a car driving, interpret the animation for me.I just want more motion control using click and drag along with typing prompts.Thank you for reading this!

This would be really helpful. I like animated gifs because they will automatically play without people having to click play on the viva engage post. I see I can do this via Express so have downloaded and then reopened and re-exported in there. Would also be helpful to see the usual gif export option ie number of times it should loop but I can't see this in Express, think it must loop Forever.

BUENAS NOCHES, SOLICITO QUE LOS VIDEOS NO SEAN DE 8 SEGUNDOS, QUIERO HACER VIDEOS DE 5 MINUTOS A MAS

ElevenLabsで日本語話者を追加してほしいです。日本語を流暢に話せる話者が選択できずに制作する際に困っています。

Merhabalar, Galeri kısmında görsel arama butonu eklenebilir. Tek tek tüm sayfaları geziyorum yada varsa daha görünür bir yere ekleyebilirsiniz.  Teşekkürler

IN Firefly boards, can you add visual connectors to link content. like a flow chart, Comfy_UI etc.

see subject

I really loved the Firefly; I'm looking forward to its audio capabilities. I find that I have to use the frame to regenerate my video; there should be a provision to extend videos like Google Flow easily.

Please consider adding the marquee zoom tool to the boards

Actually, I saw that your Translate in Video system is available in many homes and our neighboring country India has Hindi language, but many of our Bangladeshi users use your Adobe Firefly, but there is a misfortune here that we are not getting our Bengali language in audio tracking or audio conversion. Please, I am making an urgent request to you to bring it soon.

figure out what a bull elk sounds like - it does not sound like a cow.  if this app can not generate a bull elk bugling, it is not ready for commercial use

add the search field in Firefly

Firefly needs a robust resolution settings feature. I would like to be able to set ppi, color mode, 8, 16 and 32 bit color, etc. 

We really need the option to use the same seed for gens, having to copy paste every time is tedious

What I propose:A Firefly feature (or prompt template) that generates visual representations of musical triads  and chords. For example:Input a chord or scale (e.g., C major, A minor)Output a creative image representing the chord: colors, shapes, patterns, or abstract visuals that match the chord’s mood.Optionally generate multiple variations for the same chord or progression.Why this is useful:Musicians and visual artists can explore music visually, blending sound concepts with creative imagery.Makes generative workflows more immersive and interdisciplinary: music theory + AI imagery.Useful for album art, social media posts, educational visuals, or inspiration boards.How it benefits Firefly community:Expands Firefly’s use beyond static image generation to musically inspired visualizations.Encourages creative experimentation across music + visual art.Fits Firefly’s goal of helping creators express ideas visually using AI.