Free image tools that run in your browser

Which tool do you actually need?

Most people arrive with a problem rather than a tool in mind. The upload form rejected the photo, or the logo has a white box around it, or the page loads slowly because every picture on it is four thousand pixels wide. Here is the shortest route from the problem to the answer.

If none of those quite matches, the search box takes plain words. Typing shrink finds the compressor, typing transparent finds the cutout tool, because it looks at what each one does rather than only what it is called.

Resizing and compressing are not the same thing

These two get muddled constantly, and knowing the difference solves most picture problems in one move.

Resizing changes how many pixels the picture contains. A photo that was 4000 pixels wide becomes 1000 pixels wide, and there are genuinely fewer pixels afterwards.

Compressing keeps every pixel but describes them less precisely, trading fine detail for file size.

Here is the part people miss. Resizing is by far the bigger lever. Halving the width also halves the height, so you remove roughly three quarters of the data in one step. No quality setting comes close to that.

So when a photo is too heavy, the first question is not how hard to squeeze it. It is whether the picture needs to be four thousand pixels wide at all. Usually it does not, because it is going into a space eight hundred pixels across and the browser was shrinking it on arrival anyway. Resize first, compress second, and do them in that order.

What size should things be?

Match the size to the space the picture will fill, and stop there. Everything beyond that is weight carried for nothing.

• Full width banner: 1920 pixels wide, rarely more.
• Image inside an article: 1200 to 1280 pixels.
• Product photo in a grid: 800 pixels.
• Thumbnail or card: 400 pixels.
• Avatar: 150 to 400 pixels square.
• Email newsletter: 600 pixels, since that is the usual column width.
• Upload form with a cap: whatever gets you under it, often 1000 or less.

Screens with high pixel density look sharper if the file has more pixels than the space it occupies, so supplying roughly double the display width is defensible when sharpness matters more than speed. Doubling everything by reflex, though, is how pages end up slow.

Choosing a format

Four formats cover almost everything, and each one is genuinely better at something.

JPEG

The safe choice for photographs. Every browser, every email client, every upload form and every ancient piece of software opens one without complaint. It cannot store transparency, and it handles sharp edges badly, which is why text in a screenshot saved as JPEG picks up a fuzzy halo.

PNG

Records every pixel exactly, so it is right for screenshots, logos, diagrams and anything with crisp text or hard edges. It also handles transparency properly, including soft edges that fade rather than stopping abruptly. It is a poor container for photographs, where it can easily be several times larger than needed.

WebP

Typically a quarter to a third smaller than a JPEG at the same visual quality, and it keeps transparency. Every current browser supports it. The friction is outside the browser, where some older desktop software and a handful of upload forms still accept JPG and PNG only.

AVIF

Smaller again, and particularly good with gradients and flat colour. Support is solid in current browsers and thinner everywhere else, so it suits a website you control rather than a file you are sending to somebody.

Why edges look ragged in some tools

There is a technical detail behind why the same photo comes out clean from one tool and speckled from another.

When an image is shrunk dramatically in a single step, most drawing code samples a scattering of the original pixels and ignores everything in between. On a brick wall or a striped shirt that produces the crawling, sparkly texture people describe as looking cheap. Fine text turns to fuzz and thin lines break up.

Reducing in halving steps instead means every pixel contributes to the result. It takes a little longer, and it is why pictures from these tools look properly resized rather than merely smaller. It matters most on exactly the images people care about: screenshots with text, product photos, anything with fine repeating pattern.

The privacy problem nobody mentions

Every photograph a phone takes carries hidden information alongside the picture: the camera model, the settings, the exact date and time, and on most phones the GPS coordinates of where you were standing.

That data travels with the file. Post a photo online and you may be publishing your home address without realising it. People have been located from a picture of a pet on a sofa.

Converting or resizing a picture here strips all of it, not as a bolted on feature but as a consequence of how it works. The image is drawn onto a canvas and written out fresh, which leaves every one of those fields behind. Running a photo through a conversion before it goes onto a listing or social media is a genuinely useful habit.

The flip side is worth stating too. If you wanted to keep the shooting date for your own records, it will not survive. Keep the original somewhere safe when that matters.

Two things that catch people out

HEIC photos from an iPhone

Modern iPhones save in HEIC by default, and browsers have never included a decoder for it. The files look like ordinary photographs in your gallery, which is why this surprises so many people. In your camera settings choose Most Compatible instead of High Efficiency and new photos will be JPEGs. For pictures you already have, sharing or exporting them usually converts them on the way out.

Transparency turning white

JPEG has no transparent pixels at all. Convert a cutout logo to JPEG and every see through area becomes solid, usually white. Most converters do this silently and you find out after placing the file. If you need transparency kept, save as PNG or WebP. If you need JPEG for a form that demands it, at least choose the colour that goes behind rather than being given one.

A sensible order of operations

When a picture needs more than one thing done to it, the order changes the result.

1. Remove the background first, since that decides whether you need transparency at all.
2. Resize next, because dropping the dimensions removes far more weight than any format change or quality setting.
3. Convert third, once you know what the picture contains and where it is going.
4. Compress last, if it is still too heavy after all that.

Doing it the other way round means you carefully compressed pixels you were about to throw away.

Working through a batch

Most of these tools take a folder at a time, and a few habits make it painless.

• Work on copies. Resizing and compressing are one directional. The pixels you remove are not coming back.
• Test on the worst one first. Find the picture with the largest area of plain sky, wall or skin. If that survives your settings, everything else will.
• Do not compress twice. Each pass discards what the last one kept and the losses accumulate. Go back to the original and use a lower setting instead.
• Check the result at full size. A thumbnail hides everything. Open the saved file properly before sending a batch you cannot easily redo.

Why none of this leaves your device

Think about what actually goes through an image tool. Scanned identity documents for a visa application. Photographs of children. Medical letters. Product shots before launch. Screenshots of internal dashboards. The ordinary route is to upload all of it to a free website found through a search result and wait for a link.

Your files then sit on hardware you cannot inspect, for a period set by a policy you have not read, handled by a company you knew nothing about five minutes earlier. There is a particular irony in uploading a photograph to a stranger’s server in order to strip the GPS data out of it.

Every tool here uses your browser’s own image decoder and writes the new files in the same tab. Nothing is sent anywhere. Open the network panel in your developer tools and watch while it runs if you would rather see that than take our word for it. Once a page has loaded it keeps working with the connection off entirely, and there is nothing to delete afterwards because nothing was ever stored.

The limits, stated plainly

Working locally has one real constraint: your own device does the work, so its memory sets the ceiling.

A laptop handles large batches of high resolution photographs comfortably. A phone manages smaller ones and will struggle with forty forty megapixel images at once. Where a tool has a sensible cap it tells you the number and the reason rather than failing quietly, and splitting a very large batch in half solves it almost every time.

In exchange there is no account, no daily allowance, no queue, no watermark and no size limit sold back as a subscription.

Beyond pictures

Images are one part of the toolkit. If your photographs are heading into a document, the PDF tools handle merging, splitting and compressing. The text and speech tools read the words out of a screenshot or a scan. And if you are putting an application together rather than just a file, the career tools cover resumes, letters and letterheads.