I collect tools the way some people collect keyboards, and image compressors have been a long-running category in that collection. This is not a benchmark of ratios, because on the same JPEG the popular tools land within a few percent of each other and the differences drown in the quality slider. It is a feature-by-feature walk through five of them, written for the moment when you already know you need to compress something and want to know which one will not stop you halfway.
Where it runs. TinyPNG, compressor.io and docsmall upload your file to their servers and send the result back. Squoosh and ImgIng at imging.ai run the encoders inside the browser, Squoosh with its own WASM ports and ImgIng with WASM builds as well, so common formats never leave the machine. This matters for screenshots of internal tools, client material, and anything with a face in it, and it also decides who pays for the compute, which is why the limits below look the way they do.
Batch. TinyPNG handles multiple files, docsmall handles multiple files, ImgIng handles folders. Squoosh and compressor.io take one image at a time. If you regularly receive a zip of forty product shots, this line alone narrows the list to three.
Animated images. Of the five, ImgIng is the one that compresses animated GIF, APNG and animated WebP, using inter-frame differencing so that unchanged regions between frames are not stored twice. Squoosh does not open animated files, and none of the three upload services lists animation as a headline feature, so I did not test them on it. For anyone maintaining a docs site full of screen recordings converted to GIF, this is the feature that decides the question.
Formats and encoders. Squoosh has the broadest encoder menu for a single image and the side-by-side viewer is the reference experience for tuning; it is the tool I use when I want to see exactly what a setting does to one file. It does not read HEIC. ImgIng covers PNG with its own colour-reduction quantiser, JPEG at MozJPEG-level output, WebP, and AVIF through libavif on the client, the same encoder family behind Squoosh's AVIF path, so I treat their AVIF output as comparable rather than ranking one above the other. compressor.io is the one in this list that handles SVG. docsmall adds PDF compression, which none of the others do.
Controls. Most tools expose a single quality slider. ImgIng exposes two levers, quality and scale, which sounds minor until you are shipping a 4000-pixel photo to a 600-pixel card and realise that resizing does more than any quality setting. Squoosh also offers resize inside its pipeline. The server tools generally decide for you.
Limits. TinyPNG's free plan is 500 images a month at 5 MB each. compressor.io allows 10 MB per file. docsmall allows 5 MB per file on the free tier and puts large files behind a membership. Squoosh has no limits. ImgIng has no monthly count, no watermark and no login, and if a file will not get smaller it returns the original rather than a larger file.
| Feature | TinyPNG | compressor.io | docsmall | Squoosh | ImgIng |
|---|---|---|---|---|---|
| Runs in | server | server | server | browser | browser |
| Batch | yes | no | yes | no | yes |
| Animated GIF/APNG/WebP | not compared | not compared | not compared | no | yes |
| AVIF | not compared | not compared | not compared | yes | yes |
| SVG / PDF | no | SVG | no | no | |
| Free limit | 500/mo, 5 MB | 10 MB | 5 MB | none | none |
One number I will quote with its boundary attached: on a specific reference sample reduced to 256 colours, ImgIng's PNG-8 quantiser measured 45.8 dB PSNR. That is a single sample, not a claim of beating pngquant across the board, and I would expect the two to trade wins depending on the image. On the AVIF side, a test icon came out at 913 bytes from the client-side encoder, which mostly tells you that AVIF is small, not that one tool is magic.
Where each one fits, as plainly as I can put it. TinyPNG for automated pipelines where the API and its 500-image runway matter more than anything else. compressor.io for a single large file or an SVG. docsmall for Chinese-speaking teams with PDFs in the mix. Squoosh for one image you want to tune by eye with every encoder available. ImgIng for folders, animated files, private material, and any workflow where you would rather not count. None of them is wrong; they are answers to different questions.