Compress a PDF to 100 KB (and other hard limits)
How to hit an exact size target, and what every method costs you in quality.
Why 100 KB is hard
A 100 KB budget for a multi-page PDF is brutal. One clean phone photo alone is typically 300–800 KB. To fit, pages must become heavily-compressed images — which is why "compress to 100 KB" is always a quality decision, not just a size one. Tools that promise otherwise are quietly downsampling to near-illegibility anyway.
The reliable recipe
- Cut pages first. If a form allows 5 pages and you need 3, split out the pages you need — halving pages beats halving quality.
- Compress with a stated target. Use a tool that takes the size as input and searches for the best quality that fits — like the target-size compressor here. Quality-slider tools make you guess-and-retry five times.
- Check readability. Open the result at 100% zoom. If small print in scans became mush, raise the target or reduce page count instead.
Typical limits you'll meet
| System | Common limit |
|---|---|
| Government tender portals | 100–500 KB per document |
| University submission portals | 200 KB – 2 MB |
| Legacy e-mail servers | 5–10 MB total |
| Visa / immigration portals | 240 KB–4 MB, varies by field |
What compression actually does
All web compressors fundamentally re-encode page images. Text becomes part of the picture (no longer selectable) and fine detail softens. Two things separate good implementations from bad ones: whether the quality search is systematic (binary search beats fixed "low/medium/high"), and whether the original stays untouched on your disk. Both are true for the tool above — and nothing is uploaded to anyone's server.
If text must stay selectable
Then image-compression is the wrong tool. Re-export from the source (Word, LaTeX, InDesign) with downsampling disabled, embed only the fonts you use, and delete embedded debug or attachment data. Vector text is nearly free in size; images are where your kilobytes actually live.