To compress an image, re-encode it at a lower quality setting, or convert it to a more efficient format like WebP. At quality 80, the default worth trusting, a typical phone photo comes back 60 to 80 percent smaller with no change you can see. A browser compressor does this in seconds, nothing to install, no signup.
That is the short answer, and for most files it is the whole answer. The rest of this guide is the full map of how to compress an image on anything you own: what compression actually does to the file, the browser route with the real numbers, the built-in tools on Windows, Mac, and Office, phones, formats, and what to do when a form demands an exact size in KB. I built the compressor on this site, so every number here is measured, not quoted.
Here is the routing table up front. Every row is expanded in a section below.
| Your situation | Do this |
|---|---|
| One photo over an upload limit | Browser compressor at quality 80, pull the slider lower if needed |
| A folder full of photos | Batch compress, up to 30 at a time free |
| Images for a web page | Resize to display width, compress, serve WebP |
| Screenshot, logo, chart, or text | Keep it lossless, use PNG or lossless WebP |
| No internet, or a file you will not upload | Preview on a Mac, Paint on Windows |
| A bloated PowerPoint or Word file | Compress Pictures inside the Office app |
What does compressing an image actually do?
Compressing an image re-encodes it so the same picture takes fewer bytes, and there are exactly two ways to do that: lossy, which discards detail, and lossless, which does not. Everything else in this guide follows from that split.
What lossy compression does
Lossy compression shrinks the file by throwing away visual data your eye barely registers. It rounds fine detail to coarser values, a step called quantization, and it usually stores color at lower resolution than brightness, called chroma subsampling, because human vision is much sharper on brightness than on color. JPEG works this way, and so do WebP and AVIF in their default modes. The discarded data is gone for good, which is why the order of operations matters later.
What lossless compression does
Lossless compression keeps every pixel exactly and saves space by describing each pixel as a difference from its neighbors instead of spelling it out in full. PNG is lossless, and WebP and AVIF both offer lossless modes as well. The file shrinks less, but the image that comes out is identical to the one that went in. Google's web.dev performance guide is the best neutral reference on both mechanisms, and it matches what I see in the encoder every day (web.dev on image performance).
Which type fits which image?
Which kind you want depends on the content, not on preference. Photographs are full of noise and soft gradients, exactly the material lossy compression eats without a trace, so photos shrink dramatically at no visible cost. Text, line art, charts, and sharp-edged logos are the opposite case: lossy encoding smears their edges.
A screenshot of a spreadsheet wants PNG. A picture of your dog does not.
How to compress an image in the browser, step by step
The fastest way to compress an image online is a browser compressor, because there is nothing to install and the defaults are already right. I will use the one on this site for the walkthrough because I built it and can vouch for its numbers byte for byte.
- Open the free compressor on the home page and drop your image on it. You can drop up to 30 images at once, up to 25 MB each, with no signup.
- Leave the quality at 80. That is the level the web has settled on as visually lossless for photographs.
- Read the after size on the result card. If you need the file smaller, pull the quality slider down and watch the number drop until it fits. Re-compressing the same upload at a new quality or format never counts against you.
- Download the result. The original on your device is untouched, and the uploaded copy is deleted from our storage within the hour.
What happens to your file on the way through
Two useful things happen on the way through that you would otherwise do by hand. EXIF data, the hidden metadata your camera writes into every file, including GPS coordinates and the exact moment you took the shot, is stripped by default. And JPEG output is written progressive, so on a slow connection it renders in passes instead of loading top to bottom.
The measured before and after results
Here is what that pipeline did to two real files in July 2026, measured, not estimated.
| File | Before | After (quality 80) | Result |
|---|---|---|---|
| Phone photo, JPEG | 4.5 MB | ~900 KB JPEG | ~80% smaller |
| Home page demo, JPG | 497 KB | 214 KB WebP | 56.9% smaller |
The phone photo landed at the top of the typical 60 to 80 percent range because camera JPEGs are saved at a quality far above what any screen shows. The demo file landed just under that range, and its win came mostly from the WebP conversion rather than from a big drop in quality. Both patterns repeat all day long in the logs.
Compressing vs resizing: which one do you need?
Compressing changes how the bytes are stored; resizing changes how many pixels exist. They both make files smaller, they are constantly confused, and picking the wrong one is the most common way people ruin an image, so it is worth thirty seconds to separate them.
When to resize an image
Resizing, properly called resampling, removes pixels. A 4000 by 3000 photo resampled to 2000 by 1500 has a quarter of the pixels and lands somewhere near a quarter of the size, before compression even starts. Do this when the image is physically larger than anywhere it will ever be shown, which describes almost every camera photo headed for a web page or a chat thread.
One warning: dragging a corner handle in a document changes the display size only. The file underneath keeps every byte.
When to compress an image instead
Compressing keeps the pixel count and re-encodes the data more aggressively. Do this when the dimensions are already right and the bytes are not, which is the usual case for upload limits.
The right order: resize first, then compress
The best results come from doing both in the right order: resize to the dimensions the destination actually displays, then compress the result. A photo resized to half its width and compressed at quality 80 will be a fraction of the size of the same photo run through either step alone, and it will look identical on the page. The two halves of that workflow have their own deep guides: make an image smaller: bytes vs pixels untangles the two meanings of smaller, and reduce image resolution covers the pixel side.
How much can you compress before it looks bad?
Quality settings run on a 1 to 100 scale and the trade is direct: the lower the number, the smaller the file and the more detail is discarded. The useful range is much narrower than the scale suggests, and knowing where the cliff is saves a lot of second guessing.
Where image quality visibly breaks: 80, 60, 50
At 80, a photograph is visually lossless at normal viewing size, which is why it is the default here. At 60, the file comes back far smaller again and most photos still hold up on a screen. Below about 50, the compression starts to show its hand, first as blocking in smooth areas like skies, then as softness in skin and fine texture. When I need a hard target I start at 80 and step down, checking the preview at each stop rather than trusting the number.
The recompression trap
The mistake that actually wrecks images is not choosing 60 instead of 80. It is compressing a file that was already compressed, saving it, then compressing the result again next month, stacking losses generation after generation. Lossy compression is a one-way door. Compress once, from the original, sized for the destination, and keep that original somewhere safe.
If the destination demands a tiny file and you are worried about what will survive, the pillar on compressing without losing quality maps where the real limits are, format by format.
Which format makes the smallest file: JPEG, PNG, WebP or AVIF?
For photographs, AVIF is usually smallest, WebP is close behind, and JPEG is the compatibility baseline; for graphics with sharp edges, lossless WebP beats PNG. Sometimes the biggest saving is not the quality slider at all but the format the file is saved in.
How much smaller are WebP and AVIF files?
The numbers are consistent enough to plan around. Lossy WebP comes out 25 to 35 percent smaller than a JPEG of the same visual quality, and lossless WebP is around 26 percent smaller than the equivalent PNG, per Mozilla's format documentation (MDN image file type guide). AVIF often lands around 50 percent smaller than JPEG, at the cost of slower encoding. The demo measurement above is the concrete case: the same picture as WebP was 56.9 percent smaller than its JPG source.
| Format | Compression | Typical saving | Best for |
|---|---|---|---|
| JPEG | Lossy | baseline | Photos going anywhere old or unknown |
| PNG | Lossless | larger than JPEG | Screenshots, logos, text, transparency |
| WebP | Both | 25 to 35% under JPEG | The modern default for the web |
| AVIF | Both | ~50% under JPEG | Maximum savings where support is known |
When to convert the format, and when not to
The tool converts between JPG, PNG, WebP, and AVIF on output, so switching format is a dropdown, not a separate errand. One honest caveat: AVIF support is still not universal, so if the image is going somewhere you do not control, WebP is the safer modern choice and a JPEG fallback for old clients is not paranoia. And in the other direction, never save a photograph as PNG. It is the most reliable way to make a photo file several times bigger while it looks exactly the same.
Compress an image on Windows, Mac, or in Office
Every desktop you own can shrink an image without installing anything, and the built-in routes matter for the day you are offline or holding a file you would rather not upload. The catch is that most of them work by removing pixels rather than by re-encoding at a chosen quality, so know which lever you are pulling.
Compress an image on a Mac with Preview
Open the image in Preview, choose Tools, then Adjust Size, tick Resample image, and enter smaller dimensions. The dialog shows the resulting file size before you commit, which Apple documents in its Preview guide (Apple Support on resizing in Preview).
Compress an image on Windows with Paint
Open the image in Paint, click Resize on the Home tab, switch the units to pixels, enter smaller dimensions, and save as JPEG. Like the Preview route, this is a resize-first method, fine when losing pixels is acceptable.
Compress pictures in Word, PowerPoint, and Excel
Office is its own case, because a ten-slide deck full of phone photos balloons past what email will carry. PowerPoint, Word, and Excel share a Compress Pictures command under Picture Format. The default target is 220 ppi, good for screens and most printing, and ticking Delete cropped areas of pictures discards the parts of images you cropped away but the file still secretly carries (Microsoft's current Office guidance).
A note on what you will find when you search this topic: the instruction page Google's AI answer box most often cites is written for Microsoft Picture Manager, a product discontinued years ago (Microsoft's legacy compress-a-picture page). If the steps you found mention Picture Tools that no longer exist, that is why. The methods above reflect current software.
For the complete device-by-device walkthroughs in one place, the pillar on compressing images on any device goes OS by OS.
Can you compress an image on your phone?
Yes, and you do not need an app for it. The browser compressor runs in the phone's browser exactly as it does on a desktop: open the site, tap to pick photos from your camera roll, and download the compressed copies back to the phone. Same engine, same quality 80 default, same metadata stripping, and the GPS coordinates your phone camera embeds in every shot are exactly the metadata most worth stripping before a photo leaves your hands.
Why phone photos are so big in the first place
Phones are where compression matters most, because that is where the oversized files are born. A modern phone camera saves at a quality meant to survive poster-size printing, which is why every photo is four to six megabytes and why most upload forms choke on them. The phone does not know where the photo is headed. You do, and that knowledge is the entire trick: compress for the destination.
Native options on iPhone and Android
Both major platforms also have native tricks, from share-sheet size options to resizing by screenshot, and they behave differently enough that they earned separate spoke guides. The device pillar linked above covers them in full, iPhone and Android both, with measured results for each route. One habit worth keeping: compress on the phone before you share, because then every app downstream gets the small file and nothing recompresses it into mush later.
What if you need an exact size in KB?
Upload forms with hard limits, 200 KB for a passport photo, 100 KB for a job portal, 50 KB for some government sites, are the one case where you compress to a number instead of to a look. The workflow that gets you there without trial-and-error frustration has three steps.
Step 1: resize the image before you compress it
A form that wants 100 KB almost always wants a small display size too, and a 4000 pixel wide photo will not reach 100 KB at any quality worth having. Cut the dimensions to roughly what the form displays, and the byte target becomes easy instead of impossible.
Step 2: work the quality slider against the live after size
Drop the quality, read the number, stop when you are under the limit. On this tool, re-running the same image at a different quality or format costs nothing, so the loop takes seconds.
Step 3: sanity-check the arithmetic
The measured phone photo above went from 4.5 MB to about 900 KB at quality 80, a factor of five. Getting that file to 100 KB is another factor of nine, which no quality setting delivers alone but a resize to a third of the width, roughly a ninth of the pixels, does. That is the worked example to keep in your head: dimensions buy you orders of magnitude, quality buys you the last factor of two.
The upload-focused pillar, compress images for email, web, docs and uploads, carries the per-destination limits and recipes.
Where should you go next?
This page is the hub for the whole compression topic, and each spoke below goes deeper than a hub should. Pick the one that matches the problem in front of you.
- Reduce image file size: every method is the master list of byte-cutting techniques with worked examples.
- Make an image smaller: bytes vs pixels untangles the two meanings of smaller, the confusion behind most bad results.
- Reduce image resolution covers the pixel side: when to resample and to what dimensions.
- Compress multiple images at once solves the folder-of-photos problem in batches instead of one at a time.
The sibling pillars cover the neighboring territory: any device for OS-by-OS routes, uploads for destination limits, and quality for keeping every detail that matters.