How to use Expand Image
Upload the image to expand
Add it under Your image: JPG, PNG or WebP up to 10 MB. Edges that are easy to continue, such as sky, walls, floor or sand, extend best. A subject cut off by the frame or text at the border is harder to rebuild.
Say what fills the new space
In "What should appear in the new space?", name what continues and what must not appear. For example: "more beach and sky on both sides, no new people". A Style preset such as Wider scene or Add headroom replaces the text with a ready instruction. There is no format control: the result keeps the supported shape closest to your upload.
Pick a model
Choose one under Model. The models differ in how literally they follow the instruction and how photoreal the new area looks. If the extension misses, try the same text on another model before rewriting it.
Check the seams, then adjust
Press "Expand image" and look where old and new areas meet. Lines, light and perspective should continue without repeated patterns. The original part should still match your upload. If the new area drifts, ask for a smaller extension.
What a real job costs here
The cheapest model on this page charges 14 credits and the most expensive 38 credits. At the best pack rate that is about $0.1490 up to $0.4044.
| Pack | Credits | Runs with GPT Image 2 (medium) |
|---|---|---|
| $9 | 750 | 53 |
| $19 | 1,650 | 117 |
| $39 | 3,500 | 250 |
| $79 | 7,300 | 521 |
| $149 | 14,000 | 1,000 |
Choosing between the models on this page
GPT Image 2 (medium): Posters, menus, interfaces and any image with readable text. Typography that stays readable at small sizes. Watch out for: not the most photoreal option.
FLUX 2 [max]: Everyday photoreal image generation and iteration. Photorealistic default. Watch out for: text rendering is weaker than GPT Image 2.
GPT Image 2.5: Edits and product shots where the subject must stay recognizable. Sharper detail and more natural light. Watch out for: small, dense text can still break.
Nano Banana Pro: Hero images where realism decides the result. Photoreal skin and texture. Watch out for: higher rate than everyday image models.
The crop you regret
Every archive has a shot that is right but framed too tight. Outpainting rebuilds what was outside the frame instead of cropping into what is left.
The shape stays close to your upload
There is no format control here. The result comes back in the supported shape closest to your photo. That means a portrait photo does not turn into a 16:9 banner. The subject gets smaller and more of the surroundings appear.
Say what belongs there
Left alone, the model continues the scene. Name what should appear and you get more control over the new space.
Frequently asked questions
How far can it expand?
Moderate extensions look best. Very large jumps invent more and drift further from the original.
Does the original stay untouched?
Not pixel for pixel. The model redraws the whole frame, so compare the original part with your upload: small details can shift.
What does it cost?
Per image, by model: GPT Image 2 (medium) 14 credits, FLUX 2 [max] 16, Nano Banana Pro 38. The exact total appears above the button before you generate.
Is image expand the same thing?
Yes. Image expand, outpainting and uncropping all mean growing the frame beyond the original edges.
