How to compress an image
Choose a JPG, PNG or WebP from your device. Enter the maximum size requested by the destination form in the target field. This tool uses decimal kilobytes: 200 KB means 200,000 bytes. That number is a usage example, not a current requirement from a particular organisation. Check the destination’s accepted formats and minimum dimensions before preparing your final file.
Maximum width and height define a bounding box. The image keeps its proportions and is never enlarged. Compression first searches for a suitable encoding quality at the fitted dimensions. If additional resizing is enabled, it can reduce the dimensions further to approach the byte target. The final report includes the resulting width and height so you can check that the image remains large enough.
File size and pixel dimensions are different
Pixel dimensions describe the number of image samples; file size describes the bytes needed to store them. A simple graphic and a detailed photograph can have very different file sizes even at identical dimensions. Reducing width alone therefore cannot promise a particular KB result. This tool measures the actual encoded output instead of presenting an estimated download size.
When the tested outputs cannot meet the target, the tool says so. You can still download the smallest tested output, but that does not establish whether an upload service will accept it. Disable additional resizing if your pixel dimensions must remain fixed. For a demanding target, you may need to allow a larger file or choose an image with fewer details.
Choosing JPG or WebP
JPG is a practical output for many photograph workflows. WebP provides another lossy encoding option; verify that the destination accepts its extension. Both outputs in this tool use a white background. Transparent areas in a PNG become white, including when WebP is selected. Keep that in mind for logos, interface assets and other artwork that depends on transparency.
Re-encoding does not promise to preserve camera metadata, colour profiles or other file properties. Animated inputs are treated as a single image; the output does not retain animation. Browser encoders and colour handling can produce different results on different devices. Where product colours, small lettering or facial details matter, open the downloaded file in its destination application as well.
Reviewing the result
Compare the original and output previews alongside the file size and resolution report. The preview is not an automatic quality certification. Look closely at text edges, faces and fine textures. If the result looks unsuitable, raise the target size and try again. The download uses a different filename so that you can keep the original image for another attempt.
For a website, start with the space where the image will appear. Serving a very large source in a small card can transfer unnecessary bytes. If you need to design a new social sharing card, use the OG image generator. Compressing one image does not resolve every performance problem: loading behaviour, layout and other page resources also affect the experience.
Privacy and practical limits
The browser opens the source, draws it to a canvas and encodes the result locally. The image content is not sent to a processing server. The share action never embeds the image in the link. You can optionally share the size and format settings; the recipient must choose their own file. Manage the downloaded image yourself when using a shared computer.
This version accepts one file up to 20 MiB and a decoded image up to 32 million pixels. These are application limits chosen to contain browser memory use. A compressed file can require substantially more memory after decoding. If processing fails on a phone, try a smaller source image. SVG, HEIC and PDF are outside the supported input formats for this version.
Frequently asked questions
Can I produce a file of exactly 200 KB?
The tool attempts to produce a file at or below the chosen byte ceiling. It does not pad the output to an exact size. Check the destination’s format and dimension rules as well as the final number of bytes.
Is my photo uploaded?
Image processing happens inside your browser. A shared settings link contains no photo. Downloading saves the result to your device; sending that file to an application or website is a separate action you control.
Can I download the result as PNG?
This version accepts PNG input but exports JPG or WebP. Both output options flatten the image onto white. Transparency and animation are not preserved, so keep the source if those properties are important.
Why is the output still above my target?
Detailed images and strict resolution requirements can make a small byte target unattainable. Enable additional resizing or increase the target. The report always shows the measured output size rather than labelling an oversized file as successful.