Question 1

What is outpainting?

Accepted Answer

Outpainting is the opposite of inpainting. Inpainting fills in missing regions inside an image (e.g. removing a watermark); outpainting extends the image outward, generating plausible content beyond the original borders. Both use the same mask-based AI editing — only the mask geometry differs.

Question 2

How much can I extend an image?

Accepted Answer

We offer +25%, +50%, and +100% extensions in each chosen direction. Smaller extensions (+25%) give the most reliable results because the AI has more surrounding context to work from. +100% is dramatic but requires the AI to imagine more, so results can drift away from the original style.

Question 3

Will my original image be preserved exactly?

Accepted Answer

Yes. The mask tells gpt-image-1 to only regenerate pixels outside the original image region — the original area is preserved bit-for-bit. The "seam" is where the AI has to blend its new content into your existing edge, and our prompt is specifically tuned to make that seam invisible.

Question 4

What works best for outpainting?

Accepted Answer

Images with continuous backgrounds (skies, landscapes, water, walls, skin) outpaint beautifully because the AI only needs to continue existing textures. Images with complex structured content (furniture, faces near the edge, text) are harder because the AI has to guess what else was in the frame.

Question 5

Is this the same as just adding padding in Photoshop?

Accepted Answer

No. Photoshop padding adds empty space or clone-stamps neighboring pixels. Outpainting actually generates plausible new content — continuing a beach, adding more sky, extending a room, revealing more of a face that was cropped. The AI reasons about what should be there based on visual context.

Question 6

What image sizes does this support?

Accepted Answer

Input can be any size up to 10MB. The output is clamped to gpt-image-1's max dimensions (1536×1024 or 1024×1536). If your composed canvas would exceed that, we proportionally downscale the composition — the ratios stay correct, just at a smaller final resolution. Run our AI Upscaler afterwards if you need more pixels.

Question 7

Can I outpaint the same image multiple times?

Accepted Answer

Yes. If +50% isn't enough, download the result, upload it again, and outpaint once more. You can chain runs to eventually get very wide or very tall compositions. Each run counts against your daily quota.

Question 8

Why did my result look weird at the edges?

Accepted Answer

Two common causes: (1) the original image has important content right at the edge with no blank space, making the seam hard for the AI to blend; (2) you asked for too large an extension on a complex scene. Try a smaller scale (+25%) for your first run and see if the result looks clean, then iterate from there.

把任意图像向外扩展

Programmatic mask, pixel-perfect edges

All directions supported

Seamless edges

Flexible scale

Pay only for what you use

Three steps to a wider frame

Upload the image

Pick direction and scale

Click Expand Image

常见问题解答

想要更精细的控制？