Compress Image Online
Shrink photos for faster websites, lighter emails and quicker uploads.
How this compressor spends bytes
Compress Image does not zip the file or delete hidden tags. It decodes the picture into a bitmap, then asks the browser to encode that bitmap again as JPEG or WebP. JPEG and WebP both throw away detail the eye is less likely to notice: smooth sky, soft shadows, and fine noise go first. Edges and text survive longer, which is why a portrait at 75% can look fine on a phone while a screenshot of a spreadsheet at the same setting turns fuzzy around the letters.
The quality slider runs from 20 to 95 and starts at 75. That number is the encoder quality passed to canvas.toBlob, not a promise about file size. A noisy night photo at 75% can still be large. A flat graphic at 75% can collapse. The preview waits about 220 milliseconds after you stop moving the slider, then re-encodes so the note can show the original size, the new size, and a savings percent.
JPEG has no alpha channel. If the source has transparency, the JPEG path paints the canvas white before drawing, so clear pixels become white. WebP keeps the pixels the browser encoder can store; it does not paint that white fill. PNG is not an output option here. If you need a lossless PNG, use Convert Image. If the goal is to drop GPS and camera tags, compression is the wrong tool: a new JPEG often loses EXIF because the container is rebuilt, but that is a side effect. Use Remove Metadata when privacy is the job.
| Control | What the page actually does |
|---|---|
| Formats | Output is JPG or WebP only. The browser must be able to decode the file you drop in. |
| Quality | 20–95, default 75. Sent as a 0.20–0.95 quality argument. |
| JPEG background | Transparent pixels are flattened onto #ffffff. |
| Savings label | Percent is floored at 0. Compare the two byte counts in the note if the file grows. |
| Download name | Original base name plus -compressed.jpg or -compressed.webp. |
| Network | The photo stays in the tab. Nothing is posted to ImageSwift for this tool. |
Why a “smaller” setting can make a bigger file
Re-encoding is not a ratchet. If the original is already a harsh JPEG at a low quality, encoding it again at 90% cannot restore the lost detail, and the new file can be larger because the second encoder describes the blocky image less efficiently. The status note will not print a negative percent. It prints 0% smaller and still shows both sizes, so read the kilobyte figures, not only the percent.
Doing this twice on the same download is worse. Each pass quantizes again. Keep the camera original, and treat every compressed file as a delivery copy for a specific slot: email, product thumbnail, or article body.
Jobs this slider is built for
- A product photo that must sit under a shop upload limit. Start at 75% JPEG, then step down only while the edges of the object stay clean.
- A blog hero that will be shown around 1600 pixels wide. Compress after you resize. Shrinking the pixel grid removes more bytes than the quality slider alone.
- A chat screenshot you will send once. WebP at 75% is often enough, and you do not need a print master.
- A logo with flat color and sharp type. This tool is a poor fit. Lossy photo codecs smear edges. Convert to PNG instead and, if the pixel size is huge, resize first.
A practical pass
- Drop one image. The Compress button stays disabled until a file is loaded.
- Leave quality at 75 and look at the size note after the preview settles.
- Switch between JPG and WebP. WebP is often smaller at a similar look, but some older mail clients still want JPEG.
- If the picture is a photo of type or a UI, move the slider up before you download. If it is a soft photo for a thumbnail, you can go lower.
- Download
-compressed.jpgor-compressed.webp. Do not feed that file back into the tool for a second pass.
Frequently asked questions
Does Compress Image upload the photo?
No. Decoding, the white fill for JPEG, and the new encode all run in your browser. The file is not sent to an ImageSwift server for this tool.
What does the quality number mean?
It is the encoder quality from 20 to 95, starting at 75. Higher keeps more detail and usually a larger file. It is not a target file size, and two different pictures at 75% will not land on the same number of kilobytes.
Why did the file get larger?
A high-quality re-encode of an already small or already damaged JPEG can add bytes. The savings percent is never shown below zero, so compare the two sizes in the preview note and pick the smaller file.
Can I compress a PNG and keep transparency?
Not on this page. JPEG output paints transparency white. WebP is the other output, and PNG is not offered. Use Convert Image when you need a PNG.
Will this remove GPS coordinates?
Do not count on it. The new file is a fresh encode, so many camera tags disappear, but this page is not a metadata editor. Remove Metadata rebuilds the image for that purpose and names it with -clean.
Should I resize before I compress?
Yes, when the display size is smaller than the camera file. A 4000-pixel photo encoded at 75% is still a large image. Resize to the width you will actually show, then compress the result.