Everything you need for images

Which image tool do you need?

Format, size and dimensions are three different problems, and reaching for the wrong one is the usual reason a file comes out bigger, blurrier or the wrong shape.

The file is too large to upload

Compress an image keeps the format and the dimensions and trades quality for size. If the picture is also far larger than it needs to be on screen, resizing it will save more than compression ever can — a 4000px photo shown at 800px is carrying sixteen times the pixels it needs.

Something will not accept the format

Older desktop and business software often predates WebP and HEIC. WebP to JPG and HEIC to JPG exist for exactly that, and the batch converter moves between JPEG, PNG and WebP when you have a folder rather than a file.

The image is the wrong shape or has the wrong thing in it

Crop chooses which part of the picture to keep; resize changes how many pixels the whole picture has. Cropping to a platform aspect ratio and then resizing to its pixel size is the order that avoids surprises.

You need the subject without the background

Remove the background produces a transparent PNG, or puts the subject on a plain studio color. This is the one tool on the site that downloads something: its segmentation model and runtime arrive from the model host the first time you use it. Your image is not among those requests.

The photo should not carry where it was taken

Phone photos store GPS coordinates, the camera body, and the date and time in EXIF. Remove EXIF Data shows you what is actually in the file before it strips it, which is more useful than a tool that silently removes something you never saw.

The pictures should be one document

JPG to PDF and PNG to PDF assemble images into a single PDF with control over page size and orientation — the usual answer for receipts, scans and screenshots that have to be sent as one file.

What these tools will not do

These tools decode an image, work on the pixels, and encode a new one. Everything below follows from that.

• Quality lost to JPEG does not come back. Converting a JPG to PNG preserves the pixels you currently have, losslessly, but it cannot reconstruct detail the original encoder discarded. The PNG will usually be several times larger for no visible gain.
• Transparency is not created, and it can be destroyed. JPEG has no alpha channel, so PNG to JPG composites transparent areas onto white and the flattening is permanent. Going the other way does not add an alpha channel to a photo that never had one.
• Smaller format does not mean smaller file. WebP usually beats JPEG at the same visual quality, but not always, and re-encoding an already-optimized image can make it grow. The tools report both sizes so you can see rather than assume.
• Metadata is not carried across a conversion. Color profiles and EXIF are generally not copied into the output. That is helpful when you wanted them gone and surprising when you wanted them kept.
• Background removal is a model, not a person. Fine hair, glass, motion blur and low contrast between subject and background are where it struggles, and it will not ask before guessing.
• Your device sets the ceiling. Large batches, very high-resolution photos and first-run model downloads all use real memory, and a phone has less of it than the machine a tool was tested on.

Where the processing happens, and how to verify it yourself in about a minute, is set out on Security and File Processing.