Nano Banana 2.1 · 16:9 · 2K · 120 credits
Architectural fashion
A fashion editorial of a woman in a green dress walking down a bright yellow staircase against a pink motel facade. Geometric composition, warm sunlight.
Type what you want to see and get a PNG picture up to 4K in the shape you pick, made by Nano Banana 2.1 or GPT Image 2.
One subject and one clear style per prompt works best.
0 / 5000
90 credits for this image
Describe what you want to make.
No watermark · Image download · Credits back if a run fails

Architectural fashion
Nano Banana 2.1 · 16:9 · 2K
A fashion editorial of a woman in a green dress walking down a bright yellow staircase against a pink motel facade. Geometric composition, warm sunlight.
Every picture below came from the exact prompt under it, with no retouching. Each card shows the model, shape, size and credits, so you know the cost before you write your own.
Swipe to inspect all 6 prompts
Editorial demo. Results vary between runs.
Text to image AI creates a new picture from a written description, with no camera, stock photo or drawing skills. Wikipedia defines a text-to-image (T2I) model as one that "takes an input natural language prompt and produces an image matching that description."
People search "text to img" or "AI picture generator from text" for four distinct jobs. Use this text-to-image studio for a new PNG or quote graphic; switch to image to image AI to edit a photo.

Render "WAVE FORM" as metallic sculptures on black, directional studio light.
| You want to… | You give it | You get | Where |
|---|---|---|---|
| Create a new picture | A prompt | A brand-new image | This page |
| Turn a quote into a graphic | Exact words plus a style | Your words, designed | This page, GPT Image 2 |
| Read words from a photo | A photo | Editable text | An OCR tool |
| Change your own photo | A photo plus an instruction | Your photo, edited | Image to image AI |
Summary: words go in, a new PNG picture comes out. To edit a photo you already have, use image to image.
Write a prompt, pick a model, shape and size, then generate and download a PNG. The button shows the credit cost before you run it.

Name subject, action, place, framing and style, or click Use this prompt on any card.

Nano Banana 2.1 is default. Pick a shape, then 1K, 2K or 4K.

Press Generate and save a clean PNG. Change one detail and rerun anytime.
| Spec | Nano Banana 2.1 | GPT Image 2 |
|---|---|---|
| Output | 1 PNG per run | 1 PNG per run |
| Sizes | 1K, 2K, 4K | 1K (up to 1,536 px), 2K (up to 2,496 px) |
| Shapes | 8, from 1:1 to 21:9 | 5: 1:1, 3:2, 2:3, 16:9, 9:16 |
| Prompt length | 5,000 chars | 4,000 chars |
| Credits per image | 90 / 120 / 250 | 120 / 350 |
Specification summary: one PNG per run, up to 4K with Nano Banana 2.1 or 2K with GPT Image 2, from 40 credits per picture.
Describe the picture in this order: subject, action, place, composition, style. Google's Nano Banana prompting guide uses the same formula and says to "describe what you want, not what you don't want".
Our text to image generator follows clear order: subject, action, setting, camera angle, then lighting or material cues.

Woman in green dress on yellow stairs, pink facade, warm light.
Put words in "quotes", give font, size and position, and ask once. OpenAI's GPT Image prompting guide advises this plus spelling rare brands letter by letter.

Kraft coffee poster, headline "SINGLE ORIGIN" at top.
| Part | Example |
|---|---|
| Subject | A corgi in a yellow raincoat |
| Action | Jumping over a puddle |
| Place | A wet Tokyo street at dusk |
| Composition | Low-angle close-up |
| Style | 35mm film photo, neon reflections |
| Words (optional) | Headline "RAIN DAY" in bold white sans-serif, top centre |
Summary: stack prompts from subject to action, place, composition, and style, placing any quoted headline last.
| Weak | Strong |
|---|---|
| cute dog photo | A corgi in a yellow raincoat jumps over a puddle on a wet Tokyo street at dusk, low-angle close-up, 35mm film photo. |
| a cool coffee poster | A launch poster for a coffee roaster, kraft paper texture, headline "SINGLE ORIGIN" at the top. No other text. |
Summary: concrete subjects, lighting, camera framing, and quoted words turn vague prompts into repeatable pictures.
Pick Nano Banana 2.1 for detailed photos, wide shapes and 4K; pick GPT Image 2 when the picture must carry readable words or an exact layout.
1K / 2K / 4K · 90+ cr
Photo-real scenes, products, food and wide banners.
1K / 2K · Quoted Type
Posters, labels, character sheets and readable copy.
Keep Subject Look
Restyle uploaded photos or hold one character steady.
720p / 1080p MP4
Turn finished stills into clips with native sound.
| Feature | Nano Banana 2.1 | GPT Image 2 |
|---|---|---|
| Best at | Photo-real scenes, products, food, banners | Posters, labels, character sheets, text |
| Top size | 4K | 2K |
| Credits at 1K / 2K | 90 / 120 | 120 / 350 |
Model summary: Nano Banana 2.1 is cheaper and goes to 4K; GPT Image 2 costs more at 2K (350 vs 120 credits) but is stronger for text. More on the Nano Banana 2.1 image generator and GPT Image 2 image generator.
Anyone who needs one specific picture today and can't shoot it switches models and ratios right in the browser.
1:1 at 1K for 90 credits in this text to image generator, then animate it with image to video AI.
Corgi in raincoat jumping a Tokyo puddle, 35mm close-up.

Need the same character in many pictures? Make the first one here, then keep its look with image to image AI.
Text to image models still slip on long text, exact counts and keeping one character identical across runs. Even Google's Imagen, Wikipedia notes), "has difficulty rendering human fingers, text".
Text-only prompts re-roll faces each run and export flat PNGs. Draft one sheet here, then lock identity in image to image.

| Limit | What happens | Do this instead |
|---|---|---|
| Long or tiny text | Misspelled letters | GPT Image 2, short text in quotes |
| Same character twice | Faces and outfits drift | Image to image with a reference |
| Exact counts, hands | Six apples instead of seven | Ask for fewer items |
| Real people, logos | Not allowed by our terms | Describe a fictional person |
| Layers or vectors | One flat PNG | Use a design tool |
Limits summary: the AI is strongest for one striking picture; for exact copy or a repeated character, pair it with image to image.
One picture costs 40 to 350 credits. With the $9.90 pack of 1,500 credits that is $0.59 to $2.31 per image, and as little as $0.22 on a monthly plan.
| Plan | Price | Credits | Nano Banana 1K images | GPT Image 2 2K images | Per 1K image |
|---|---|---|---|---|---|
| Small pack | $9.90 once | 1,500 | 16 | 4 | $0.59 |
| Medium pack | $24.90 once | 4,000 | 44 | 11 | $0.56 |
| Large pack | $49.90 once | 9,000 | 100 | 25 | $0.50 |
| Starter | $14.90 / month | 3,000 | 33 | 8 | $0.45 |
| Pro | $24.90 / month | 9,000 | 100 | 25 | $0.25 |
| Premium | $99 / month | 40,000 | 444 | 114 | $0.22 |
Price summary: $0.22 to $0.59 per 1K picture; pack credits never expire; 4K needs pack credits or Pro or Premium, not Starter.
Failed or undersized pictures return their credits automatically; ones you dislike do not. See our refund policy for the 7-day pack refund.
Because you can see how each picture is made and what you may do with it.
1K–4K
Your prompt goes word for word to Google's Nano Banana 2.1 or OpenAI's GPT Image 2, verified to paid dimensions with zero recompression.
30 Days
We don't train AI models on your prompts or pictures. Files delete after 30 days; see privacy policy.
0 Logos
Use PNGs commercially under terms. See U.S. Copyright Office and Instagram label notes.
Auto
Credits are held at Generate and spent only on delivery.
Edit it, animate it or try another model with the same credits.
Edit & RestyleEdit or combine up to 0 photos.
Animate StillTurn your picture into a 2–15 second clip with sound.
All ModelsEvery image model in one place.
Prompt to ClipDescribe a moving clip instead.
Text to image in one prompt: describe it, see the price, get the picture.