Bug Reports

I am genuinely disgusted by the quality of Adobe Firefly's user experience.For a company that has built its reputation on professional creative software, Adobe has managed to deliver an image-management system so poorly conceived, so frustrating, and so unreliable that I seriously question whether anyone responsible for designing it has ever tried using it with a large collection of generated images.Adobe Firefly is a paid creative service, yet its fundamental image-management functions are practically unusable.Let me explain exactly why.1. Generation History: An Endless Warehouse of ImagesEvery image you generate goes into something Adobe calls "Generation History."Sounds reasonable, doesn't it?Except this supposedly professional image library becomes an absolute nightmare when you've generated hundreds or thousands of images.Images take forever to load. Sometimes they never load at all. The screen fills with grey placeholders, and users are left endlessly scrolling and refreshing, ho

Les demandes de modification de la voix ne sont pas prises en compte. Il s’agit de remplacer la voix d’une femme par celle d’un garçon (qui existe). J’ai beau redemander (10 générations, 2000 crédits), rien n’y fait. Je déconseille formellement Firefly video 2 qui n’est pas du tout mûr pour être commercialisé. 

In the recent update to the Firefly interface, “Reuse Image Settings” has been removed. I use this ALL the time. Please put it back. Or tell me where it is if it’s hidden somewhere.--UPDATE: There is now an icon (two arrows in a circle) under “Generations” only (not Files) that will let you reuse settings.

We use Firefly's Expand (Crop and expand) to put product photos on exact square canvas sizes for a rental catalog. After the redesign this stopped working. The Upload button on the 'Edit an image' screen opens only the system file picker, which cannot be used by automation or assistive tools. 'Edit image' on an existing result gives 'Can't load item.' In the prompt bar, the X to remove the 'current canvas image' reference does nothing when clicked. I could not find any way to switch back to the previous interface. Please bring back the previous editor (or a toggle to it), or add a way to drop an image straight onto the Expand canvas. This is currently blocking our work. Thank you.

I'm reporting a sudden and significant increase in generative credit costs affecting the entire GPT Image 2.5 model family in Adobe Firefly, including both Sunburst and Flare.1. What happened?On November 5, 2026, generating a 2K image using GPT Image 2.5 Sunburst at High quality cost 15 generative credits.Today, November 6, the exact same configuration suddenly costs 40 credits — an increase of approximately 167% overnight.After investigating further, I discovered that this issue is not limited to Sunburst. GPT Image 2.5 Flare is showing the same unusually high credit costs. The issue affects all available resolutions (1K, 2K, and 4K) and quality settings.2. Current credit costsThe following prices (in generative credits per image) are displayed for BOTH GPT Image 2.5 Sunburst and GPT Image 2.5 Flare:Quality 1K 2K 4K Low 5 5 10 Medium 15 20 30 High 30 40 60 Extra High 40 70 100 Max 70 100 130 For example: 2K High increased from 15 to 40 credits

in image to video mode, I should see the option for first frame and last frame upload, for me only first frame shows up. This is the case for all available models 

I input my description and reference images, and this message keeps popping up when I try to generate an image.  What do I do?414 Error - The Request could not be satisfied 

чому не дозволяє генерувати коли оплата на місяць?

Since September 26, every partner model in Firefly fails on my account: Gemini 3.1 (Nano Banana 2), Nano Banana Pro, GPT Image, Flux, Imagen. Adobe's own Firefly models work fine. The UI shows "Something went wrong. Please try again later." Under the hood it's POST firefly-3p.ff.adobe.io/v2/3p-images/generate-async -> HTTP 408 {"error_code":"timeout_error","message":"system under load"}, returned in 0.5-2 seconds every single time. That's not a load timeout, it looks like the account is blocked from partner models. Same on the web app and in Photoshop, different browsers and devices. Latest test: Nov 6, 2026 14:17 UTC, Request ID 18b39c79-cfde-47d3-b618-a66e46dc9489. What support did so far (case ADB-47459129-Y0V2):- Oct 8: asked for HAR, sent Oct 9 with request IDs- Oct 12: "passed to Engineering"- Oct 20: confirmed it's a known bug on Adobe's side affecting multiple users, closed the case, suggested I buy a second plan on another email as a workaround- Nov 2: my reply bounced to a

i am using kling 3.0 via adobe firefly to generate 9:16 videos . but as soon as i click generate it shows “something is wrong” . i have tried it with multiple prompts .

For about a week, every image generation in Firefly AI Assistant fails with: "The generation tool is temporarily unavailable." Model: Nano Banana Pro.Sometimes Firefly doesn't load at all, just a spinning circle. The sign-in verification code window also didn't load.What I tried:- Safari, Chrome and Opera- Two different computers- Wi-Fi and mobile data- Logging out and back inSo it is not my computer or internet. Please check my account. I use Firefly for my business and can't work. Screenshot attached. 

On Kling 3.0 OMIN, my elements are no longer working. The timeline keeps prompting me to create new ones, and they are not generating properly, with errors occurring when I try to use older elements.

The Adobe MCP connector can start spoken-content summaries but provides no usable way to retrieve their results. Please route this to the Adobe connector/MCP integration team if this forum is not the owning team; the connector's Support link led to the Firefly server, which directed bug reports here.Environment: Codex desktop on macOS; Adobe connector plugin version 10.0.0. Observed November 6, 2026.Steps to reproduce:1. Upload a video with narration using asset_upload_file_from_url. Upload succeeds and returns an ACP asset URN.2. Call media_summarize with that asset URN.3. The response returns status "working", a UUID-format taskId, and adobeTaskPoll.pollTool = "summarizePoll" (poll interval 15 seconds).4. Inspect the available Adobe tools: summarizePoll is not exposed. The available image_get_task_status tool says to use it for any working taskId.5. Pass the returned taskId unchanged to image_get_task_status.Actual result:Parameters failed connector schema validation: taskId [pattern

Hi Adobe Support Team,I've been trying to generate videos using Firefly and its available models since yesterday, but I keep getting error messages and haven't been able to generate anything.I've already cleared my cache, deleted old videos, and confirmed that my internet connection is working properly. I also have more than enough credits available.Could you please look into what's causing this issue? It's been happening since yesterday, and it's starting to delay my work.I'd really appreciate your help in resolving this as soon as possible.Thank you!

when I want to use generative fill, it takes too much time, and kind a freeze the options, when finally let me generate de changes I want, do not let me see to easy, I need to zoom in or zoom out to let me see the options of the changes

Third-party video generation models in Adobe Firefly are failing to start.I have tested the same simple text-only prompt with no uploaded reference images across multiple models.Working:Adobe Firefly Video ModelNot working:Kling 3Kling 3 OmniGoogle VeoGoogle Veo FastGeminiRunwayWhen I press Generate with any of the third-party models, the Generate button/interface briefly flickers but no generation starts. There is no progress indicator and no video is produced.The exact same simple prompt generates successfully when I switch back to Adobe Firefly Video.I have reproduced the problem in Safari, Chrome and Opera, so it does not appear to be browser-specific.My account shows 10,010 premium generative credits available.Please check whether third-party/partner video model access is correctly provisioned on my account, or whether there is currently a backend routing/entitlement issue affecting partner models.Reproduction steps:Open Generate Video in Firefly. Select a third-party model such a

I used to work last week. Does not work at all anymore. 100% of the time does not work. Always "something went wrong, Please try again later"

Hello Adobe Firefly Team, I’ve been experiencing issues with Firefly while generating images. The tool is not producing accurate or proper visuals as per the given prompts — the output often looks distorted or unrelated to the described design. I’ve tried rephrasing the prompts and using different modes (Text to Image and Generative Recolor), but the issue persists. Please look into this or suggest how to resolve it.Thank you for your support.

Ingreso en generar vídeos indico que es lo que requiero y la app solo me indica que hubo un error que intente más tarde 

this is the output of that generation request

Ich kann keine Bilder mehr generieren. Es kommt immer die Meldung :Da hat etwas nicht geklapptVersuche es später noch einmal. oder  Sorry, unsere Server schmelzen gerade.Zu viele Personen generieren gerade – bitte warte, wir haben bald wieder freie Kapazitäten.

It displays the thumbnail after I enter the prompt then just before I click download the page just bombs out and takes the credit off me. This is happening continually - MAC OS Tahoe. 

While working on a professional project using Adobe Firefly Video (Veo 3.1) with reference image and custom text prompt, the generation failed twice consecutively showing the "Impossibile caricare" error message. Despite the rendering failure and no output video being produced, 400 generative credits were deducted per attempt (total 800 credits lost). I request a review of this rendering failure and a credit refund to my account. 

I’ve created the video with an animated avatar and a script. I can see that the avatar is saying the script but there is no audio - I can’t hear the avatar speaking. I moved the video over to clipchamp but it doesn’t show that any audio came across with the video.