Image Compressor
Compress JPG, PNG and WebP in your browser with a quality slider (1-100) and a max-width cap. Files never leave your device.
FAQ
Does compression noticeably reduce quality?
It depends on the setting. Between 60 and 80 quality is hard to tell apart from the original in most photos. If you need to get under a hard limit such as 100KB, also reduce the pixel width instead of only lowering quality.
Can I compress an image to under 100KB?
Yes. Lower the quality slider and reduce the max width — the before and after sizes are shown side by side. For forms and exam uploads with strict limits, adjust both.
Are my images uploaded?
No. Compression runs in your browser with Canvas, so the files never leave your device and it works offline.
Which formats are supported?
JPG, PNG and WebP as input. Output can stay in the original format or be converted to JPEG, PNG or WebP.
About this tool
The pipeline is fixed: read the file as a DataURL, decode it into an image, redraw it at the target size on a Canvas, then re-encode with canvas.toBlob at the chosen format and quality. All of it happens in the browser — nothing is uploaded and it works offline, so ID scans and contract photos are safe to drop in here.
Three controls do different jobs: quality (1–100, mapped to an encoder value of 0.05–1), max width (downscales proportionally when exceeded) and output format. To cut size meaningfully, shrinking pixels beats lowering quality: going from quality 90 to 70 saves a little, while going from 4000px to 1600px wide can halve the file. Set a max width first, then tune quality — the result looks far better than dragging quality down alone.
Two format facts worth knowing: the quality slider does nothing when the output is PNG (PNG is lossless and has no quality parameter), so to shrink a PNG you must reduce dimensions or switch to JPEG / WebP. And JPEG output is painted onto a white background first, so transparent areas become white — that is what "JPEG has no alpha channel" means in practice, not a bug. Keep transparency by choosing WebP or staying with PNG.
The percentage in the result compares against the original file bytes, so "larger by xx%" happens normally: an already-compressed source, or a conversion to lossless PNG, can only grow. Re-encoding also drops the EXIF in the original (capture time, camera model, GPS coordinates) — good for privacy, bad if you wanted to keep copyright metadata.