Image Compressor (Free, No Upload — Reduce Image File Size)

A free tool that compresses image files. Just drag and drop your JPEG, PNG or WebP and it cuts the file size while keeping the pixel dimensions unchanged. Everything runs inside your browser and no image is ever sent to a server. Compress several files at once and download them together as a ZIP.

How the formats differ

Format	Compression	Transparency	Characteristics
JPEG	Lossy	No	The best fit for photographs. Lowering the quality shrinks the file, but the discarded detail can never be recovered.
PNG	Lossless	Yes	Suited to illustrations and screenshots. Being lossless, lowering quality does nothing, so real savings require moving to a lossy format.
WebP	Both	Yes	Roughly 20–30% smaller than JPEG at the same quality, and it supports transparency, so it also replaces PNG well.
GIF	Lossless	Yes	Limited to 256 colours. Poor for photographs; for a still image, moving to WebP gives a much smaller file.

Choosing a quality value

Quality	How it looks and when to use it
90–100	Loss is essentially invisible. Good for photos destined for print or for material you will edit later. The saving is modest.
75–89	The practical sweet spot. For a website or a social post, almost nobody spots the difference from the original.
60–74	Noise starts to show around edges when you zoom in. Perfectly adequate for thumbnails and gallery listings.
30–59	The loss is clearly visible. Keep this as a last resort for extremely tight size limits.

This tool keeps the pixel dimensions (width × height) and saves space through compression alone. So if you want to bring a 4000 px wide photo down to 1200 px for a website, change the dimensions first with our image resizer and then compress — the result will be far smaller.

What image compression does

Compressing an image means shrinking its file size while keeping it looking as close to the original as possible. A photo from a modern phone often runs to 3–5 MB; put it straight onto a website and pages crawl, attach it to an email and you hit the size cap. Because a barely perceptible drop in quality can cut the file to a fraction of its size, it pays off whether you are publishing, sharing or archiving.

This tool keeps the pixel dimensions exactly as they are and reduces size purely by adjusting the lossy compression quality. Everything happens inside your own browser, so no image is ever sent to a server. Most online compressors upload your file first; here you can safely process family photos and internal screenshots that should never leave your machine.

How to compress an image

Choose your images Drag them onto the drop zone or click to browse. Several files can be processed together.
Pick an output format If in doubt, leave it on Automatic. PNG cannot be made smaller as PNG, so Automatic converts it to WebP, which also keeps transparency.
Set the quality Around 80 is a good starting point. Watch the reduction figure and lower it if you need more.
Download the results Save files individually, or grab everything at once as a ZIP.

Tips for getting more out of it

Look at dimensions first, quality second. If a 4000 px wide photo is going on a website, resizing it to around 1200 px with our image resizer cuts the file to nearly a tenth on its own. That beats relying on compression alone.
A photo saved as PNG will not get smaller if you keep it as PNG — the format is lossless, so there is no quality to give up. Convert it to WebP or JPEG instead (Automatic picks the right one).
Re-compressing an image that was already heavily compressed can make it larger. When that happens the row shows "+x% (larger)", so simply keep your original.
WebP is 20–30% smaller than JPEG at matching quality. Every major browser supports it, though JPEG remains the safer choice if the file is going into older software.
Logos and icons that need transparency will get a white background if you convert them to JPEG. Choose WebP to keep the alpha channel while still saving space.

When an image compressor helps

Speeding up a website

Images account for most of a page's weight. Compressing before upload makes pages load faster, which helps both bounce rate and search performance.

Meeting a form size limit

Application forms and submission systems often cap each image at 2 MB. You can meet the limit without a visible drop in quality.

Emailing a batch of photos

Attach a few dozen photos untouched and you blow past the limit instantly. Compress them together and a single message will usually do.

Tidying up phone storage

Dropping the quality slightly for archive copies frees up a surprising amount of space without deleting any memories.

Image compression terms

Lossless compression: Compression from which the original data can be restored exactly. PNG and GIF work this way: no quality is lost, but the savings are limited.
Lossy compression: Compression that throws away detail people are unlikely to notice in order to save far more space. JPEG and WebP work this way, and the discarded data is gone for good.
Quality: A 0–100 value setting how much detail lossy compression retains. Lower numbers mean smaller files and more noise around edges.
WebP: An image format developed by Google. At matching quality it is 20–30% smaller than JPEG, and it supports transparency and animation.
Alpha channel: Per-pixel transparency information. PNG and WebP can carry it; JPEG cannot, which is why transparent areas end up filled in.
Mosquito noise: The shimmering speckle that appears around text and edges when lossy compression is pushed too hard.

Frequently asked questions

They are never even uploaded. Reading and compressing both happen inside your browser, and the image never travels across the network. That makes the tool safe for family photos and internal documents alike.

Around 80 is the usual answer. For a website or social post, almost nobody notices the difference from the original at that level. Use 90 or above for print and for material you will edit later, and 60–74 for thumbnails.

Because PNG is lossless — there is no quality to trade away. To make it smaller here, convert to WebP, which keeps transparency, or to JPEG for photographs. Automatic picks the appropriate one for you.

That happens when an already heavily compressed image is re-encoded, or when you export at a higher quality than the original was saved at. The row will show "+x% (larger)" — just keep your original file.

No. This tool keeps the pixel dimensions and saves space through compression alone. To change the actual width and height, use our image resizer. Resizing first and then compressing gives a far smaller file.

Side Note — how JPEG exploits the weaknesses of human sight

JPEG was standardised in 1992, and its underlying idea was less about packing data efficiently than about discarding what people will not notice. The image is cut into 8×8 pixel blocks, each block is re-expressed as a stack of fine striped patterns, and precision is then stripped away starting with the finest stripes — the ones the eye is least likely to catch. When you drag a quality slider, what you are really choosing is how much to throw away.

The design leans cleverly on how our vision works. Human eyes are sharp at detecting changes in brightness but far duller at changes in colour. JPEG exploits that by often recording the colour information at half resolution. For a fixed amount of data, spending more of it on brightness simply looks better — a classic trick that dates back to the early days of television broadcasting.

What is curious is that this discarding behaviour means a JPEG degrades every time it is saved again. Open one, change nothing, save it, and tiny errors still accumulate through the compress-and-expand cycle. It is why images reposted around the internet acquire that distinctive grit, and the effect became a genre of its own in the "deep-fried meme".

WebP, by contrast, was built by Google in 2010 by carrying techniques from the VP8 video codec over to still images. Video saves space by reusing the parts that neighbouring frames have in common; WebP applies that same predictive machinery within a single image, guessing the contents of a block from its neighbours and recording only the difference. The 20–30% edge over JPEG at equal quality comes from exactly this kind of accumulated hindsight.

Browse all trivia