Text to Image AI
Write who or what, where, and the look. Pick a shape, press Generate, and the picture is back in seconds.
Start from a recipe
Each picture came from the words on its card. One tap puts the words, the model and the shape in the generator.
PosterPaper crane over the city
A Moroccan girl of about ten riding a giant paper crane cinematic poster look over a night city full of lights
A 3D avatar of you
A woman in her thirties, ginger bun, freckles, smiling soft 3D cartoon plain peach background
WallpaperSnowy window wallpaper
A mug of cocoa and fairy lights on the sill vertical wallpaper, no people a snowy cabin window at night
ProductLip balm listing shot
A round tin of lip balm with no label product photo on pink sand with a dried rose
BlogBlog header: home office
A desk by a window with a laptop, a plant and a mug morning light a tidy home office, no people
Thumbnail: the giant pumpkin
A man with a curly beard looking shocked at a giant pumpkin bright thumbnail style on a kitchen table
How the image generator works
Words in, a picture out. Nothing to upload.
Describe the picture
Who or what is in it, where, and the look. One or two sentences.

Pick the shape
Square for a profile, tall for a phone, wide for a header or a thumbnail.

Get the picture
Back in seconds. Change one word and run it again to compare.
One idea, the right shape
Pick the shape for where the picture is going, then write for it.
Profile picture 1:1
Square, for avatars, feeds and app icons.

Phone wallpaper 9:16
Full screen on a phone, a story or a reel cover.

Shop listing 3:4
The shape most shops and marketplaces show first.
Thumbnail 16:9
Wide, for YouTube, slides and blog headers.
SGirl and craneCNight cityAPoster lookPCity lightsWhat a good picture prompt is made of
Four parts. The first one is the only one you must write; each of the others removes a guess.
“A Moroccan girl of about ten with curly hair riding a giant paper crane, over a night city, cinematic poster look, full of lights.”
Which model for which picture
GPT Image 2.5 Flare is the default. Switch when the picture needs a face kept, words drawn or a poster finish.
| Your picture | Pick | Why | Credits | |
|---|---|---|---|---|
Any idea, first try | GPT Image 2.5 Flare | Follows long, detailed instructions closely | 3 · 1K | |
| Nano Banana 2.1 | Fast all-rounder. Best for keeping a face the same | 4 · 1K | ||
| Nano Banana Pro | The most detail and the most accurate lettering | 6 · 1K | ||
Layouts and diagrams | Qwen Image 3 | Clean layouts, lettering and diagrams | 8 · 1K | |
Posters, big prints | Seedream 5.0 Pro | Bold colour, posters and sharp large prints | 9 · 1K | |
Looks like a photo | FLUX 3 Klein | Photo-real pictures and edits, up to 4K | 3 · 1K |
Then make it move
The picture from the first recipe sits in the corner; the clip made from it on Image to Video plays behind. Hover or hold the picture to see it whole.
Seedance 2 · 720P 5sAfter the picture
Keep the one you like and take it further.
Your picture
Download it at full size, ready to post or print.
Fix one thing
Change the colour, the background or the outfit without starting over.
Image to ImageAI image generator FAQ
What people ask before their first picture.
How does an AI image generator work?
You describe the picture in words: who or what is in it, where, and the look. The model draws a new picture to match, usually in a few seconds. Change a word and run it again to compare.
Is the AI image generator free?
A new account gets 10 free credits, and the cheapest picture is 2 credits, so you can make a few before you pay. Packs start at $9.90 and plans at $9.9 a month. A run that fails gives its credits back.
Which model should I use?
GPT Image 2.5 Flare is the default at 3 credits and follows long prompts closely. There are 11 in all; the table above shows which one suits faces, lettering, posters or photo-real pictures.
Can it put words in the picture?
Yes. Write the exact words in quotes, such as a sign that says “Grand Opening”. Nano Banana Pro gets lettering right most often; keep it to a few words.
Can I get the same character in several pictures?
Describe them the same way every time, with age, hair and clothes. For an exact match, make the first picture, then add it as a photo on Image to Image with Nano Banana 2.1 and change only the scene.
What sizes and shapes can I make?
Square, tall and wide, from 1:1 to 16:9 and 9:16, at 1K by default and up to 4K on some models. Pick the shape before you generate; it changes how the scene is laid out.
Can I use the pictures for my business?
Yes. Pictures made with paid credits carry no watermark and can go on your shop, ads, posts and print.
Other ways to start
Same credits, a photo to start from or a moving result.
ModeHave a photo or a sketch already? Change it instead.
ModeThe same kind of words, but a moving shot with sound.
ModeBring your new picture, or any photo, to life.
ModePut a person and a product together in one clip.
Say it, see it.
Your words go straight into the generator above.
