Photo Studio: make any picture from your words

The tabs at the top of the box pick how you start. Describe it makes any picture from your words. Your product: add a photo of your product (drop it in or press to choose) and the product stays exactly as it is, its shape, colors, label and words, while the scene around it is new. Photo Studio looks at the photo and suggests 10 scenes to press (clean studio shots, lifestyle, premium ad looks, seasonal), or type your own. Restyle a photo: add any photo and it keeps the layout and subject while the look changes completely; pick styles below, press one of the 10 suggested looks, or describe it. Both save your photo to your Graphics Library in a folder called Photo Studio uploads.

On Describe it, Add inspiration pictures takes up to 3 pictures whose look, colors and mood guide the new picture (not their exact content). Press the × on one to take it off.

The box at the top left takes any description, as long or short as you like ("a cozy bakery window at dawn with fresh croissants and a cat asleep on the sill"). Press the microphone to say it instead. Press Ctrl+Enter (Cmd+Enter on a Mac) to make it without reaching for the button. Your own words always come first: the looks you pick are added after them.

Under the box, the looks you picked show as chips. Press a chip to take that look off, Surprise me for a random mix (a style plus a few others), or Clear looks to start over. With an Industry scene picked, you can leave the box empty: the scene's own example subject is used.

The looks panel has a tab for each kind of look, each with real sample pictures made from the same subject, so you can compare them fairly: Style (56 in groups: Photo, such as Real photo, Ultra-real, Cinematic and Magazine editorial; Futuristic, such as Hyperreal sci-fi, Holographic, Liquid chrome and Solarpunk; Infinite Soldier, led by Everything is Possible, with cosmic ∞ light, aurora and Dawn over Earth; Positive & inspiring; Art & paint; Illustration; 3D & craft; and Creative & surreal), Medium & art movement (16, such as Charcoal, Japanese woodblock, Art Deco, Pop art and Stained glass), Lighting (26, such as Golden hour, Studio softbox, Neon glow and Backlit silhouette), Camera & shot (28, such as Portrait lens, Drone aerial, Top-down, Tilt-shift miniature and Long exposure), Color mood (24), Composition (22, including Room for words, which leaves space for a headline), Borders & frames (20, such as Instant photo border, Gallery mat, Neon frame, ∞ Infinity frame, Holographic foil and On a billboard) and Industry packs.

Industry packs are ready-made scenes for 18 kinds of business: food & restaurants, real estate (13 scenes, from bright living rooms and twilight exteriors to curb appeal, virtual staging and open houses), fashion, events & weddings, products & online stores, beauty & spa, fitness, automotive, travel & hotels, home services, health & dental, pets, construction & trades, coffee & bakery, jewelry & luxury, finance & professional, kids & education, and tech & software. Press an industry chip to see only its scenes.

Style takes up to 3 picks and Medium up to 2, blended into one look (for example Ultra-real + Neon cyberpunk); the numbers on the samples show the order. Every other tab takes one, and the tabs combine (styles, a Lighting and a Camera shot all at once). Press a picked sample again to take it off. A dot on a tab means something is picked there. The arrows at either end of the tab row show the rest of the tabs. With no style or medium picked, the picture is a real photo.

Sizes are in the Size + settings card above the looks: quick chips with their shape (Square, Portrait, Story, Widescreen, LinkedIn banner, Letter flyer) and All sizes for every one, grouped as Social (Square, Portrait 4:5, Story 9:16, Pin, Facebook cover, LinkedIn banner), Web (Widescreen 16:9, Website hero 21:9, Email header, Classic 4:3) and Print (Letter, A4, Poster 18 × 24, Postcard, Business card, Photo print). Pictures are trimmed to exactly that shape.

The same card holds every other setting. Model: Auto picks for you and says why under the card (words in quotes, or "a sign that says…", go to Words in the picture; an illustration, flat or icon look goes to Illustration; everything else, including every photo, to Everyday). Or choose one: Everyday (OpenAI GPT Image 1.5, great all-round pictures and real-looking photos), Quick & cheap (for first ideas), Words in the picture (Ideogram, for signs and posters that must spell right), Google look (Google Imagen, photo-real scenes) and Illustration (Recraft, clean designer art). Quality: Standard, or Best detail for finished work (costs more; not every model has it). See-through background: for logos, stickers and cut-outs (Everyday and Quick & cheap).

Pictures: 1 to 8 at once (2 to start). The Make button shows the total cost before you press it. Pictures appear one by one as they finish, with a line saying what is being done ("Adding golden hour lighting", "Trimming it to Widescreen").

Compare models makes the same request on four models side by side (Everyday, Words in the picture, Google look and Illustration), each labeled, so you can see which one suits the job.

Press a picture to open it big. The big view shows your words, the looks, model, size in pixels, credits used and when it was made, with the same buttons, Use these settings again, Open in your Library, and The full request sent to the model. The arrow keys move between pictures, and the address is a link to that picture you can come back to.

More like this makes new pictures in the same style, subject and mood, using the picture as the reference. Remix keeps your words and tries a new mix of looks. Change it opens a big box (with a mic): say what should change ("make the sky pink", "add a red umbrella") and press Make the change; the rest of the picture stays the same. Changes are made by the Everyday model even if another model made the picture, and the card says so.

Upscale makes a bigger, sharper copy, saved as a new picture: 2× Free runs on your own computer (the first time downloads a small free model, about 8 MB); 4× Print-sharp adds real detail on our server for a small credit; 6× Poster and 8× Big print (up to 8192 pixels on the long side) come with the paid plans, and on the Free plan they open a short note with a link to the plans.

Send to opens the picture in another tool, already placed; each choice in the menu says what happens there: Video Kit (a still clip on the timeline), Motion Maker (to animate it), Magic Canvas (a new design with the picture on it), Pro Designer (to retouch it or add text) or Mockup Studio (as the art on a product). Each of those tools then offers Add more from Library to bring in other saved pictures too. The download button saves the full-size file.

Under every picture, How did this picture turn out? takes a thumbs up or down. A thumbs down asks what went wrong; Send and fix it makes a corrected version from your words, and the report reaches the aiKit team so the tool keeps improving.

History, under your pictures, keeps every picture you made with its exact words, looks and model, newest first. Search your words to find one, or show only Made, More like this, Remixes, Changes, Upscales or Compares. Show more loads older pictures.

Press Select, then tap several pictures to send them together to Video Kit, Motion Maker or Magic Canvas, or Remove from history. Removing takes them out of Photo Studio only: they stay in your Graphics Library until you delete them there.

Each picture uses a little of your AI credit: Quick & cheap about 0.08 of a credit (0.2 at Best detail), Words in the picture 0.12 (0.36), Illustration 0.16, Everyday 0.2 (0.56), Google look 0.28. Product shots and restyles use Everyday. Suggested scenes and looks for an uploaded photo use a tiny part of a credit. Upscales: 2× free, 4× 0.08, 6× 0.12 and 8× 0.16 of a credit. The top of the page shows the credits you have left today. A picture that fails is not charged.