Compression

How to Compress an Image to 100 KB Without Ruining It

A 100 KB limit is rarely about saving bandwidth. It is a rule written into a form: a job application, a visa upload, a school portal, a marketplace listing. The form rejects the file without explaining which setting to change. This guide covers what actually moves the number, how to use the image compressor, and when 100 KB is not a fair target for the image you have.

First, agree on what 100 KB means

Most upload forms mean 100 × 1024 = 102,400 bytes. Some mean 100,000 bytes. The difference is about 2.4 percent, which is enough to fail a strict validator. Aim for 95 KB rather than exactly 100 KB so rounding in the form does not push you over.

Check whether the form also imposes pixel dimensions. Many do, and a file that passes the size check will still fail because it is 4000 pixels wide. In that case, width is the real constraint and the byte limit follows from it.

Only two settings reduce file size

For JPEG and WebP output, there are two levers:

  1. Encoder quality. Lower quality quantizes color and detail more aggressively. The file gets smaller because information is discarded, not because the data is packed better.
  2. Pixel dimensions. Halving width and height leaves one quarter of the pixels. This is the strongest lever you have, and it is the one most people skip.

For PNG there is only one lever: pixel dimensions. PNG is lossless, so a quality setting has nothing to act on. If a tool offers a quality slider for PNG, it is either ignored or it is really converting the file to another format.

Compress to a target size, step by step

The compressor works backward from the number you give it. It searches encoder quality, and if quality alone cannot reach the target, it reduces dimensions and tries again.

  1. Open the image compressor. Everything runs in the tab; no file is uploaded.
  2. Enter 95 in Target size (KB). Leave it empty only if you want a high-quality re-encode with no size guarantee.
  3. Choose an Output format. Keep the original if the destination is strict about extensions. WebP usually produces the smallest file for photographs. Choose PNG only when you need lossless output and accept that the tool will have to shrink dimensions instead.
  4. Set Max width if you already know the destination. A width of 1600 px is a sensible ceiling for an image that will only be viewed on screen. The compressor resizes before it encodes.
  5. Add files by dropping them on the page, clicking the drop zone, or pressing Ctrl+V to paste a screenshot. Batches are supported, and two files are processed at a time to keep the page responsive.
  6. Download each result, or use Download all (.zip). Every row reports the size before and after, plus the final dimensions, so you can see whether the tool had to downscale.

Files are capped at 50 MB and 40 megapixels. Those limits exist to stop a phone browser from running out of memory while decoding, not to push you toward a paid tier.

What happens when the target cannot be reached

The compressor will not silently hand you a file that misses the limit. It searches quality down to a floor of 0.3 and, if the result is still over, reduces the longest side by 15 percent and re-encodes, for up to six encode attempts in total. If the result still exceeds the target, the row says could not reach target size. At that point you have three honest options: accept a smaller pixel size, change format, or choose a different source image.

Dimensions decide whether 100 KB looks acceptable

A 12-megapixel photo compressed to 100 KB has roughly 0.067 bits available per pixel. That is not much room for a photograph with foliage, fabric, or skin texture. The encoder will produce blocking, banding, or both. The same 100 KB spent on a 2-megapixel image looks far better because each pixel gets about six times the data.

The following are rules of thumb for a typical photograph, not guarantees:

Long sidePixelsWhat 100 KB usually looks like
1600 pxabout 2 MP at 4:3Clean at normal viewing size; good for forms
1200 pxabout 1 MPAcceptable, slightly soft when zoomed
800 pxabout 0.5 MPClearly soft; fine for a thumbnail or ID photo
3000 px and up12 MP or moreVisible blocking and color banding

If a form demands 100 KB and does not specify dimensions, set Max width to 1600 and let the quality search do the rest. If the form demands both 100 KB and a large minimum dimension, the two requirements may be incompatible. Test one file before processing a batch.

PNG, screenshots, and text

The compressor treats PNG differently because it must. A browser PNG encoder produces a full-color lossless file; it does not build a reduced palette. A screenshot of a document can therefore stay stubbornly large even after the quality search, because there is no quality search to run. The tool reduces dimensions instead.

That is the wrong trade for a text document. Shrinking a page of small type until it fits 100 KB can make it unreadable. A better sequence is to convert the screenshot to JPEG or WebP first, then compress, then inspect the output at 100 percent zoom. WebP often preserves small text better than JPEG at the same byte size, but it can still soften edges. Check one page before converting a whole scan.

For more on which format to pick, see PNG vs WebP vs JPEG.

What you lose when the file is re-encoded

Browser-side encoding is a fresh encode, not a metadata-preserving edit. Expect these consequences:

  • EXIF is discarded. Camera model, exposure settings, capture date, copyright, and GPS coordinates are removed. The orientation tag is applied before encoding, so the image comes out upright even though the tag is gone. That is usually good for privacy and bad for archiving.
  • Transparency is flattened for JPEG. A PNG with an alpha channel encoded as JPEG has no alpha channel to write to. Transparent areas become a solid color, typically black. If you need a specific background, use the format converter, which lets you choose a fill color when converting to JPEG, or keep the file as PNG or WebP.
  • Repeated compression accumulates damage. Compressing an already compressed JPEG at a low quality adds another generation of loss. Start from the original whenever you can.

None of this is unique to this site. Any tool that re-encodes in the browser, and most server-side compressors, behave the same way.

A checklist before you upload

  • Target 95 KB, not 100 KB, to leave validator headroom.
  • Use WebP or JPEG for photographs; reserve PNG for images that need lossless pixels or transparency.
  • Set a maximum width based on how the image will be displayed, not on the original camera resolution.
  • Open the result at 100 percent and read any text in it.
  • Keep the original file. The compressed copy is for the form, not for your archive.
  • If the row says the target was unreachable, do not keep lowering the number. Resize first with the image resizer and compress the smaller file.

The honest summary: 100 KB is comfortable for a web-sized photograph and hostile to a full-resolution one. Decide the pixel dimensions first, and the byte target becomes easy.