Compress PDF to 100 KB

Comment utiliser Compress PDF to 100 KB

1. 1

   Drop the PDF the form rejected. The 100 KB target is already set — change it if your limit is 50, 200 or 500 KB.

2. 2

   Click Compress. The tool tries progressively stronger settings and stops at the first that fits, so quality is the best the limit allows.

3. 3

   Read the note: it tells you the final size and whether the text is still selectable.

4. 4

   Download and upload it to the portal.

Why forms ask for 100 KB, and what fits in it

The 100 KB limit is a fixture of government portals, university applications, job boards and bank onboarding forms, most of them designed when storage was expensive and never revisited. Whatever the reason, it is not negotiable, so the question becomes what can honestly fit in 100 kilobytes. Quite a lot: a one-page scanned form at 100–130 dpi in greyscale, two or three pages of a text document with a small logo, a signature page, an ID card scan. What does not fit is a full-colour, 300 dpi, ten-page scan — that one needs a split as well as a squeeze.

The mistake people make is compressing once at a fixed setting and then compressing the result again, which multiplies the artefacts without helping much. The tool instead compresses the original at increasingly aggressive settings, in one pass each, and keeps the first that lands under the limit.

How the target-size search works

First the file is rewritten losslessly — unused objects dropped, streams repacked — which alone brings some PDFs under the limit without touching a single pixel. Then the embedded images are re-encoded along a ladder: JPEG quality falls from 75 % to about 20 % while the longest edge is capped at 2000, then 1500, 1100, 800 and 600 pixels. Each rung roughly halves the image weight, and the search stops at the first rung that fits.

Only if the document has no text layer — a scan, a photographed page — does the tool go further and re-render whole pages as images at decreasing resolution. Text is never rasterized to reach a number: a page you can still select and search is worth more than a page that fits by becoming a picture. If a text PDF cannot fit, you get the smallest version found and a clear note saying so.

• Lossless rewrite first — free bytes, zero quality cost.
• Images re-encoded at five progressively stronger settings, stopping at the first that fits.
• Whole-page re-rendering only for scans without a text layer.
• The note under the result says the final size and whether text survived.

When it still will not fit

A PDF whose weight is text, vector drawings and embedded fonts cannot lose much: those things are already compact, and there is no quality knob to turn. When the note says the file is mostly text, the options are structural. Split the document and upload it in parts if the form allows several files. Delete pages that are not required — cover letters, blank backs of scanned pages, terms nobody asked for. Or print the pages that matter to a new PDF at a lower page count.

For scans, one more lever is the source: rescanning at 150 dpi in greyscale produces a file a fraction of the size of a 300 dpi colour scan, with no visible difference for a form or a signed letter.

Why this runs in your browser

Every step runs inside your browser, using the WebAssembly and Canvas APIs your device already ships with. No file is uploaded, no copy is kept, and nothing sits in a queue on a server waiting to be deleted — which is also why the tool keeps working with the network disconnected.

The PDFs that meet a 100 KB wall are the personal ones — passports, tax forms, payslips, medical letters — and the last thing they should do is sit on a stranger's server while a queue clears. Here the search runs on your own device, on the original file, with nothing sent anywhere.

Questions fréquentes

Outils associés