Take a 1024px draft to 4K without repainting it.
Generation models return a comfortable working size. Real layouts want more. This endpoint reconstructs the detail rather than stretching the pixels — and your agent can call it the moment a draft is approved.
1024 × 1024 → 2048 × 2048
Retina display assets, blog figures, in-app illustrations.
1024 × 1024 → 4096 × 4096
Full-bleed heroes, large-format screens, A4 print at 300 DPI.
1024 × 1024 → 8192 × 8192
Posters, exhibition panels, anything a viewer stands close to.
Resizing and upscaling are not the same operation
A resize has nothing new to work with. It looks at the pixels it has and guesses what sits between them, which is why an enlarged image goes soft exactly where you were hoping it would stay sharp — the edge of a logo, the serifs on a word, the weave of a fabric.
A super-resolution model has been trained on what those things look like at full size, so it can put the edge back rather than smear it. The trade is that it is a model call, not an arithmetic operation: it takes seconds and costs credits, and you run it once at the end of a pipeline rather than on every draft.
Plain resize
Free, instant, and blurrier the further you push it.
Super-resolution
15 credits, a few seconds, and detail that survives a close look.
Order of operations
Upscaling last is almost always right. Everything before it is cheaper on a smaller image.
- 1
Generate at the model’s native size
Iterate on prompts at 1024px. Nobody needs a 4K draft they are about to throw away.
- 2
Edit while it is small
Retouches and inpainting cost the same at any size, but review is faster on a light file.
- 3
Cut out the background
Matting on the smaller image gives cleaner edges and avoids reconstructing pixels you delete.
- 4
Upscale the survivor
One 15-credit call on the asset that actually ships.
- 5
Compress on the way out
Then a 1-credit compress to WebP so the 4K asset does not cost your visitors four seconds.
curl -X POST \
"https://api.imagemcpserver.com/playground/upscale" \
-H "Authorization: Bearer $IMAGEMCP_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"image": "https://cdn.example.com/draft.png",
"scale": "4x"
}'{
"name": "multicall",
"arguments": {
"requests": [
{ "tool": "upscale_image",
"args": { "image": "…/a.png", "scale": "4x" } },
{ "tool": "upscale_image",
"args": { "image": "…/b.png", "scale": "4x" } },
{ "tool": "upscale_image",
"args": { "image": "…/c.png", "scale": "2x" } }
]
}
}Both routes write one entry per image into your request log with the exact credit cost, so a batch that ran overnight is still auditable in the morning. How credits are counted.
Upscaling API questions
What scales are available?
Three: 2x, 4x and 8x. The default is 4x, which takes a 1024px generation to 4096px — enough for a full-bleed desktop hero or a decent print.
Is this different from just resizing?
Yes. Resizing interpolates between the pixels you already have, so a 4x resize is a 4x blur. Super-resolution reconstructs detail — edges stay sharp, small text stays readable, and texture does not turn to mush.
Should I upscale before or after removing the background?
After. Cutting out first means you are not paying to reconstruct detail in a background you are about to throw away, and matting is generally cleaner on the smaller image.
How much does an upscale cost?
A flat 15 credits at any scale. Upscaling runs on our own provider account, so the price is the same whether or not you have connected your own OpenRouter key.
Can I upscale a whole folder at once?
Yes. Pass the calls as a batch through multicall and they run in parallel. That is the usual shape for a nightly job that promotes the day’s approved drafts to full resolution.