How it works
The file never leaves your browser. HandyTools HQ has no upload step. pdf.js draws every page, the canvas encodes a JPEG at the quality you pick, and pdf-lib builds a new PDF of those pictures. Scanned pages with big images usually shrink. A text-only PDF can grow, because letters become pixels.
The new file is pictures. You cannot select the text afterward. If the result is larger, the page says so. Lower the quality slider for a smaller file and a softer picture.
Formula
- Approach
new PDF = one JPEG per original page - Quality
0.70 is the default; lower is smaller and softer
Worked examples
A phone scan
A 12 MB scan of a contract often drops under a few megabytes at quality 0.7, because the camera photos inside it are recompressed.
A text invoice
A 200 KB invoice made of real text may become larger. The status line compares the two sizes so you can discard the result.
| Quality | Size | Text selection |
|---|---|---|
| 0.90 | Mild shrink | Lost |
| 0.70 | Usual choice for scans | Lost |
| 0.40 | Smallest, softer type | Lost |
Common questions
Is the PDF uploaded to compress it?
No. Drawing and re-encoding happen in your browser. The file is not sent anywhere.
Why would the file get bigger?
A PDF of real text is already compact. Replacing those letters with a photo of the page adds pixels. This tool is aimed at scanned pages.
Can I still copy the text?
Not from the compressed file. The pages are pictures. Keep the original if you need to search or copy the words.
Does it change the page size?
The page stays the same size on paper. The pixels inside it are a JPEG of the drawing, at about 1.25 times the screen scale.