Compress an Image to a Target Size Without Uploading It

How it works

1. Your browser reads the image from your disk. No network request carries the picture.
2. It re-encodes the image at progressively smaller sizes and JPEG qualities until the file is under the number you picked. The first setting that fits is kept, so the result keeps as much detail as the target allows.
3. JPEG is usually the smallest output for a photograph. Choose PNG when the image has transparency or needs to stay lossless; that keeps the alpha channel but can leave the file larger.
4. If the target is out of reach, it tells you. It shows the smallest honest result and its dimensions instead of pretending it hit the number.

What it can and cannot do

A photo shrinks a lot. A 12 MP phone photo can usually reach 100–500 KB by scaling down; the result card tells you the final width and height. A screenshot or line drawing is better kept as PNG, where text and edges stay crisp.

A PNG that is really a photo usually will not reach a very small target. PNG is lossless, so a detailed photograph saved as PNG stays large even at a reduced size. Switch to JPEG and the same picture can fit easily — at the cost of transparency.

Transparent areas become white in JPEG. Keep the PNG output if the cut-out has to stay a cut-out. The tool never upscales an image, and it never returns a file larger than the one you gave it.

Metadata is not copied. EXIF, GPS location, camera model and capture time are removed because the picture is re-encoded rather than passed through. Browsers apply the orientation flag before drawing; wide-gamut photos are converted to sRGB, which can shift highly saturated colours slightly.

Animated files, SVG and some camera formats are out of scope. Animated GIF, APNG and WebP would lose frames. SVG is a vector format. HEIC depends on the browser: Safari can usually open it, while Chrome and Edge on Windows may not — export the photo as JPEG first if it fails.

Very large PNGs are capped at about 16 megapixels. Keeping the pixels, a filtered copy and the canvas in memory at once needs room; above that cap the PNG path starts from a smaller size. The result card shows the width and height it used.

The target is an upper bound, not an exact size. The tool stops at the first setting that fits, so the file is at or a little below the number you chose. We count 1 KB as 1,000 bytes, so a "100 KB" result is at most 100,000 bytes — under a form's 100 KB limit whether that form counts 1,000 or 1,024 bytes.

Pricing

Questions

Is my image really not uploaded?

Correct. Open developer tools, go to the Network tab, and compress a picture. You will see the page, the script bundle and a few tiny counter pings — no request carrying the image.

Will it hit exactly 100 KB?

It aims under, not exactly at. The first setting that fits is kept, so the download is at or below the target. That is what an upload form checks, and it avoids compressing further than the limit requires.

Which output should I choose?

JPEG for photographs, and usually for the smallest file. PNG when the image has transparency, sharp text or flat colours that must stay exact. Auto picks PNG for an image whose header says it has alpha, and JPEG otherwise.

Why is my PNG still large?

PNG is lossless, so every pixel is stored. A photograph that was saved as PNG is a poor fit for a very small target; the same picture as JPEG can be a fraction of the size. The result card shows the smallest size the PNG path could reach.

Does it remove EXIF and GPS location?

Yes. The output is a new file built from the decoded pixels, so EXIF, GPS, camera model, capture time and other metadata are not copied.

Will the result look blurry?

If the target is tight, the image is scaled down or the JPEG quality is lowered, and some detail is lost — that is what the limit requires. The result card shows the final dimensions so you can judge before downloading. Pick a larger target if the first result is too soft.

What happens to transparent areas?

PNG keeps them transparent. JPEG has no alpha channel, so transparent areas are filled with white. If the cut-out matters, choose PNG or Auto.

Can I compress an animated GIF or sticker?

No. This tool handles still images; animating output would drop frames. Use a still frame, or the Discord emoji and sticker maker for that format.

What about HEIC photos from an iPhone?

It depends on the browser. Safari can usually decode HEIC; Chrome and Edge on Windows may not. If the page says it cannot decode the file, export the photo as JPEG first and try again.

What reads the image?

The browser's own image decoder and canvas, plus a small PNG writer that runs in this tab. There is no upload and no third-party image service.

Do you store anything about me?

We count anonymous events — page opened, compression finished, upgrade clicked — with no advertising cookies, persistent analytics identifiers or image data. A strictly necessary signed cookie is used only after purchase so Pro access survives a refresh.