Image

Image is the picture side of Create. You type what you want, add photos if you have them, and the result lands in the feed above the box and in your Library.

What it does

Create turns words, and optionally your own product, person or style photos, into finished pictures. Your results sit in the feed called Your creations, and the prompt panel is docked at the bottom of the page.

Create sends your words as you wrote them. Nothing is added until you use AI Polish, Direction or a Styles pick. A small grey line under the box says so.

Your very first picture is on the Getting Started page, Your first image. This page goes through every control.

When to use it

• You want a picture of anything, in any look, from a few words.
• You have a photo of a product, a person or a look, and want a new picture that uses it.
• You want to try several ideas fast, one to four pictures at a time.

For clean marketplace pictures of one product, use Product Shoot. For a YouTube cover, use Thumbnail. For the same person in every picture, use AI Influencer from the menu. When a specialist tool is open on this page, the Create category button lists it.

Step by step

Make a picture from words

1. In the menu, open Core AI, then Image. Check that Image is selected at the top of the panel.
2. Click the prompt box. Describe the subject, the place, the light and the mood. Hindi, Hinglish or English is fine.
3. Open Shape and pick where the picture will go.
4. Set How many (1 to 4).
5. Leave Quality as it is, and Create picks for you, or pick one yourself.
6. Optional: press AI Polish for a better prompt, or open Direction to choose camera, light and look.
7. Check the number on Generate, then press it. You can keep working while it is made.

Screenshot coming soon

Make a picture from your own photos

1. Press Product, Person or Reference in the prompt box. Assets opens on the right tab: pick yours, or press Upload from device. Direction stays closed, and your pick is already selected inside it when you open it.
2. Pick a picture from the row, or press Upload in the section to open Assets, where you can also press Upload from device.
3. Type what should happen with the photo, for example "this bottle on a marble table, morning light".
4. Leave Quality as it is. A picture in a slot moves the job off Quick, because Quick works from text only.
5. Press Generate.

Change a result

1. Press the picture in Your creations to open it.
2. Under Direct this scene, press Open Direction Mode. Direction Mode opens with your picture as the first version.
3. Say what to change in plain words, such as "make it warmer". Only that part changes, and every version is kept under Versions.

Screenshot coming soon

Every option explained

Top bar of the panel

Styles, on the top line of the box, opens the Styles part of Direction, with a search box. On other tools the same button is Templates. Press a card to see What it sets and what it Adds to the prompt, then press Use style. None takes it off. A card tagged Needs your photo works only with a face photo in Person. Direction still works on top of a style.

Popular opens first. Then come the groups, including Film, Documentary, Retro, Black and white, Fun creator and India.

A style sends much more than the sentence on its card. Its full recipe goes with your words: colour grade, contrast, grain, lens, light, shadows, skin, composition, era and what to avoid, so the look holds and does not slide into a similar style. Your own words stay as you wrote them, and they still decide what is in the picture.

Upload slots

A slot takes JPG, PNG or WebP, up to 25 MB each, 16 to 8,000 pixels on a side. You can also drop a file on a slot, paste a picture into the prompt box, or drag a tile from the feed onto a slot. A picture dropped elsewhere on the panel asks What is this image? with one button per slot.

Press a filled slot and Assets opens again: your pick replaces what is there in one step. Make sheet is in Assets, on an opened product or model. The cross removes a pick, with Undo. Use a real person's photo only with their permission.

The toolbar

On a phone the box starts small, so your pictures show: one line for your words, then the shape, how many and quality button and Generate. Tap in your words and the whole box opens with the uploads, Direction and AI Polish; swipe through your pictures or press Generate and it folds again. The small arrow at the top of the box opens and folds it too. When a picture is ready it shows at the top, just above the box. Shape opens as a bottom sheet.

Quality

Auto reads your words and pictures. It moves up to Standard when you add a picture and to Refined when your picture has text on it, and the tooltip on the picker says why. Pick a tier by hand to stop Auto. Quick greys out every upload slot with the hint "Quick works from text only", and it cannot keep a locked product or person. Ultra is the sharpest, for hero shots and start frames.

Direction window

A new Direction is rolling out on Create in Image mode. If you have it, it starts with two questions.

1. What is this picture of? Person, Product, Food, Animal, Place or Others.
2. What shape? Tall, Portrait, Square or Wide. It sets the same Shape as the toolbar.

Each answer then has its own lists, in folded sections, so every option fits your picture:

Every answer also has its own Camera, Light, Look and colour and Avoid. Food has a side light and a food magazine look; a place has a drone camera and a travel photo look.

Lists follow what you pick. A cold drink shows glasses, not plates. A pet shows pets, and only what a pet can do. A home has no weather. In Others, the vehicle list shows after you pick Vehicle. Picks that cannot go together cannot both be picked: a drawn look hides the camera choices, and sun hides night.

The window runs Person, Cloth, Product, then Styles, folded, above the first question. Every tool's Direction keeps this order. Press Styles in the prompt box and it opens by itself, with two rows and Load more under them.

Realism sits just below Person, and shows only for a person.

Every image and video in a tool has an eye: in Direction, Product, Model, Assets, Prop Box, Backgrounds, sheets and every feed. With a mouse, the eye shows when you point at a tile; rest on the eye and the picture opens beside it at a large size, and a video plays muted. Clicking the tile still picks it as usual. On a phone the eye is always on the tile: press it, and tap anywhere to close. The Social feed and profile pages have no eye.

Last time brings back the last picks for the answer you chose, so your food picks stay separate from your animal picks. Magic reads your words and answers the first question for you when it is open.

Every group starts on None, and nothing is added to your words until you pick.

In this Direction, the Reference box opens Assets straight away: pick a picture or press Upload from device. It has no section in the window.

Every group starts on Auto. Pick only what matters.

Apply direction writes its words after your own in the prompt box. Applying again replaces only that part.

The feed

Your creations shows only what you made on Create, in the media of the page you are on. Chips filter by category, each with a count. Hover a tile for Add to favourites, Share to public feed, Reuse settings, Use as reference, Turn into video and Download. Press Show more for older work.

A tile in progress shows your place in line and the seconds left. The Stop button on a waiting tile takes the job out of the line and gives its credits back, and Undo puts it back. With two or more waiting you also get Stop all in line.

The opened picture

Who can use it

Free plan: included. Pro plan: included.

Your plan decides how many jobs run at once. The rest wait in line.

What it costs

Each picture is charged by its Quality tier, and the table above shows each cost. One press that makes several pictures costs the price of one times How many. AI Polish and Magic show their cost on the button when they cost credits.

A finished picture is final and paid. If a job fails or is blocked, every credit comes back and the card says how many.

Tips

• Say the subject, then the place, then the light: "a steel bottle on a wet black stone, rim light".
• Switching Image, Video or Category clears the box. Press Undo at once if you did not mean it.
• Try Quick to explore ideas, then make the best one again on Refined or Ultra.
• Put text you need in the picture inside quotation marks. Auto moves such jobs to Refined.

Common problems

"Write an idea, add an image or pick what you are making." The box is empty. Type your idea first.

The message says how many credits you need. Press Get credits.

"This goes against our content rules, so it cannot be made. No credits were used." Change the words or the picture.

"Use a JPG, PNG or WebP image." Convert the file first.

"No model here can read this many images. Remove a few." Take some pictures out of the slots.

"This style needs your photo. Add it in Person." Add a face photo to Person.

"Built for the old setup. Rebuild?" You changed a photo or choice after the prompt was written. Press Rebuild, or Keep mine.

"Taking a short break, your credits are safe." Create is paused for a moment. Try again soon.

"The model did not send a picture back, even after a few tries." The card shows how many credits came back. Press Retry.

"This is taking longer than usual. It keeps going on our side; check back in a few minutes." Your job is still running. Press Refresh.