Daily credits reminder
When your daily credits are ready, the bell in the header lets you know, and your account menu shows Collect daily credits so you can claim them in one click.
Updates
New tools, options and improvements, newest first.
When your daily credits are ready, the bell in the header lets you know, and your account menu shows Collect daily credits so you can claim them in one click.
In Review generation, each finetune now has its own strength box. Type a value from 0.1 to 1.5 to use exactly that strength, or leave it empty for Auto. You can also type a list such as "Sumi ink (image 0.8, video 0.8)" in Modify finetunes to use exactly those finetunes and strengths. Available on paid plans.
Every creator now has a public profile with a handle, bio and photo at deepany.ai/u/<handle>. Follow creators and see their new work under Showcase > Following. Make your account private under Profile > Public profile so only followers you approve see your creations. The bell in the header shows new followers, requests, comments and replies, and you can block accounts from their profile.
Studio now remembers your Remove banner choice for new tools and projects, so you no longer have to switch it on or off every time. Projects you already saved keep their own setting.
Tick Post anonymously when you comment or reply on a Showcase creation and your name stays hidden. You can still delete your own anonymous comments.
Automatic finetunes now add up to at most 1.25 strength per stage by default, which keeps stacked effects from smearing the picture. Change it in Settings under Creative fine-tuning, Finetune strength limit (0.5 to 3).
Manual Finetunes now has a single Strength setting next to Apply selection, and it remembers your choice. Click a selected finetune at the bottom to see its details; the × removes it.
The DeepAny Assistant now reads images. Attach up to 3 screenshots or results with the image button, or paste them into the question box, and ask what you need. You can also send images without typing a question.
In Review generation, Modify this prompt now also updates automatic finetunes. Ask it to remove one, add one or pick a better match, and the Finetunes list updates before you generate. Each finetune now shows its strength for every stage it applies to.
For video requests paid with credits, cancelling after you open the draft video now refunds 25% of the credits instead of all of them. Cancelling before you open it, or a request that fails, is still refunded in full. Studio asks before you open a draft and shows the exact refund in the cancel dialog.
Open any creation on the Showcase to read its comments, add your own and reply to others. Creators get a Creator badge on their comments and can remove any comment on their own creation. Comments follow the showcase rules, and you can report one for review.
Photo and Video Face Swap now count the faces in your target when you choose Swap all faces, and use one swap's credits for each face, so three faces cost three swaps. After Analyze Faces, the estimate shows the total for the faces it found. The final price always comes from the count made just before your swap starts.
While a video is generating, its progress card now shows each finished step, such as the start image, the audio and a draft of the video. Select one to open it. To turn this off, open Studio preferences and switch off Preview output stages. With it off, videos that don't pause for your choice skip the draft and finish a little sooner.
Reference to Video, Images to Video, Audios to Video and Videos to Video now offer AnyVideo 3.1 under Settings → Video model. It follows prompts more closely and handles action scenes, motion and visual styles better. AnyVideo 3.0 stays the default.
Manual Finetunes now lets you apply up to five finetunes at once. Each selected finetune has a Strength choice: Default uses its recommended strength, Dynamic eases it a little when four or five are stacked, and Custom lets you pick a strength from 0.1 to 1.5. Automatic selection still adds up to three.
The Showcase page and the Publish to Showcase window now list the showcase rules, and you confirm them before publishing. Items that break a rule are removed, and after three violations an account can no longer publish to the Showcase.
When a video request on your Unlimited pass is waiting in line and another video server is free or has a much shorter line, Studio now moves it there automatically, so it starts sooner. Requests paid with credits keep their place in line.
Open any finished image or video in Studio and choose Publish to Showcase. Your creation appears on the new Showcase page with its prompt and settings, so anyone can open the recipe in Studio and try it. Add a caption and themes, publish anonymously if you prefer, and find every creation you published on your profile, where you can remove any of them at any time. Viewers can like, search, follow a creator's page, and hide creations or creators they do not want to see. Inputs and projects stay private, and every upload is checked automatically.
Open the ShowcaseThe Subscription option on the pricing page is now called Unlimited. It works the same way: one payment for a set time, no credits used for included tools, and nothing renews. The page now explains how it works and shows the date your pass ends.
The pricing page now shows your active plans. Choose Add requests to run 2 or 3 requests at a time on the plan you already have, including lifetime plans. You pay only the difference for the time left on the plan. Extend keeps the number of requests your plan has now.
Image Edit now has an optional Edit mask input with Smart select, Draw mask, Touch up and upload, so only the area you choose changes. Turn on Auto mask in Advanced settings to build the mask from your prompt.
Inpaint Anything and Object Remover now find the area to change from your prompt when you leave Region to change empty and add no mask. Turn Auto regions off in Advanced settings to write it yourself. Guided Edit and Video Inpainting have an optional Auto mask setting that builds a Smart select mask from your prompt before generating.
Smart Select now has a Touch up button. It opens the selected mask in the brush editor, where you can paint in small missed spots or erase extra areas before saving the mask.
Cancelling a generation, deleting a project and deleting selected assets now offer Don't ask again, like the other delete questions. To get every question back, open Studio preferences, then More preferences, and choose Ask again.
The mask input in Guided Edit and Inpaint Anything now offers Smart select and Draw mask next to uploading a file. Smart select lets you describe objects or click on the image, then shows the mask before you generate. In Inpaint Anything a mask replaces the Region to change text.
Settings now has a single Quality choice: Low, Medium, High or Very high. It sets every quality option of the tool at once, including audio, motion and upscale quality for video. If you change a single option under Advanced, Quality shows Custom.
Cards for generations in progress now have an X in the top-right corner. Click it and confirm to stop that generation.
Text to Image now offers 4:5, 3:4, 2:3, 9:21 and 21:9 next to the usual shapes, and every image model makes the exact shape you pick. Choose Custom size to set your own width and height, up to about 1 megapixel.
A Servers button next to Settings opens a live view of every server for the current tool: how many jobs it runs, who is waiting, and the progress of each job. Long face-swap videos now also show when they are being split across several servers.
The Video Edit and Video Inpainting pages now play the source and edited clips on top of each other. Drag the line to compare them frame by frame.
Studio's View mode now offers Masonry, Grid and Batches. Grid lines everything up in even squares, and Batches keeps each request's results together.
Click the workflow name in the media toolbar to jump to another workflow. Input slots now open a small menu of this project's uploads, so you can reuse a file without uploading it again.
The website and Studio are now hand-translated into 19 languages. Pick yours from the globe menu in the header. Your prompts and creations are never translated.
Developers can now create with DeepAny from their own apps. Create an API key in your account, follow the developer docs, and track usage, spend and recent jobs on your API dashboard.
Read the API docsStart a new request while another one is still running. Subscriptions can include 1, 2 or 3 simultaneous requests, and you can use credits for extra ones.
See plansPlans now cover Images, Videos, Audios, Editing and Identity separately, so you only pay for the tools you use. Each plan lists the tools it includes.
Compare plansUpscale fits on one row: your source, the scale and the final size. Drop or paste an image straight into it to start.
Every verified account can now replace a character in a video with a person from a reference image.
Try Character SwapText to Image, Reference to Image and Image to Image now live together in the Image Generator. The model picker is split into 3D and 2D groups, with General as the default.
Open Image GeneratorAsk questions about tools, plans and your account from the sparkle button in the header, on the website and in Studio. It remembers your previous question, so you can ask follow-ups.
The advanced settings you knew from the original Art Generator are back in Text to Image.
Open Text to ImageImage to Video and Video Generator can keep the exact shape of your start or last frame. Turn on Keep aspect ratio, or pick Start Frame or Last Frame in the ratio list.
Open Image to VideoThe home page gallery now has 22 examples, including high-motion videos, each with the recipe used to make it.
Browse examplesWhen you change a video request and run it again within an hour, Studio reuses the stages that didn't change. You are only charged for the work that runs again.
Studio prompts can now be up to 20,000 characters long, four times the old limit, so detailed scene descriptions fit.
Upscale now has a High quality option for images and videos. It takes a little longer and restores more fine detail than Fast.