“Maximum file size: 2 MB.” “Attachment too large.” “Keep it under 500 KB.” Plenty of forms, portals and inboxes demand a PDF below a specific number, and a document that’s a hair too big simply won’t go through. The honest truth is that no tool can promise an exact size on the first try — the result depends on what’s inside your file — but you can hit almost any realistic target reliably if you know which levers to pull and in what order. This guide shows you how.
Why there’s no magic “exact size” button
A PDF’s size is dominated by the images inside it, and how far an image compresses depends on the image itself — a flat diagram squeezes far more than a busy photograph. That’s why an exact-size slider would be lying to you: the same setting produces wildly different sizes on different files. What works instead is a quick loop: compress, check the number, and if you’re still over, step up the strength or trim the document. Two or three passes gets you there. Our main guide to compressing a PDF explains the mechanics; this one is about hitting a number.
The reliable way to hit a target size
- Start with the right level. Open Compress PDF and pick a level to match your target: Recommended for anything up to a few MB, Strong when you need to get small.
- Check the result. Download and look at the file size. Close enough? Done.
- Still too big? Step up. Re-run the original on Strong rather than compressing an already-compressed file (a second pass on the same file mostly adds artefacts without shrinking much).
- Need to go further? Convert colour scans to grayscale, and remove pages you don’t need — both cut size directly.
Hitting the common targets
| Target | Best approach | Realistic for |
|---|---|---|
| Under 10 MB (email) | Recommended once | Almost any document, including scans — see compress for email |
| Under 2 MB | Recommended, or Strong for scans | Everyday reports, forms, mixed documents |
| Under 500 KB | Strong + grayscale if it’s a scan | Short documents, single-page scans |
| Under 100 KB | Strong + grayscale + remove pages | A few text-light pages only |
When the file is big for a different reason
Compression only shrinks images, so if your PDF is huge because it has hundreds of pages, no level will change the page count. Split it into parts, or extract just the pages you actually need to send — often the fastest way under a limit is to send less document, not a more-crushed one. And if a text-only PDF is already small and still “too big” for a strict field, there’s simply little left to remove; splitting is your lever.
Scans are the special case. Every scanned page is a full-resolution image, so scans compress the most — and benefit most from grayscale when the content is black text on white. If you also need the text to be searchable, run OCR first (the text layer it adds is tiny), then compress. See making a scanned PDF searchable.
Is it private, and is it free?
Compress PDF is free with no sign-up or watermark. Because re-encoding images is heavy work, it runs on our server (a “Server” badge marks these): your file travels over an encrypted connection to a hardened, network-isolated sandbox, gets compressed, and is deleted the moment it’s done — never stored or read. Grayscale runs the same way; page tools like Delete, Extract and Split run entirely in your browser.
Frequently asked questions
How do I compress a PDF to 100 KB? Use Compress PDF on Strong, convert colour scans to grayscale, and remove any pages you don’t need. 100 KB is only realistic for a few text-light pages — a big colour scan has a natural floor.
How do I get a PDF under 500 KB? Strong gets most short documents there; for a scan, grayscale it first. Check the size and step up only if needed.
Why can’t I choose the exact output size? Because how much a file shrinks depends on its images, the same setting gives different sizes on different files. A quick compress-and-check loop is more reliable than a slider that can’t keep its promise.
My PDF barely shrank — why? It’s mostly text, which is already tiny. Big savings come from image-heavy or scanned files. If it’s big because of page count, split it instead.
Should I compress the same file twice? No — re-run the stronger level on your original. A second pass on an already-compressed file adds artefacts without saving much.
Does compressing to a target lose quality? Only images are affected, and only as much as the level you pick. Text stays perfectly sharp at every level. For print, stay on Light; for screen or upload, Strong is fine.
Is it free and private? Yes — free, no account, processed in an isolated sandbox and deleted right after.
Hitting a size target isn’t about one perfect setting — it’s a short, reliable routine: compress on the right level, check the number, and if you’re still over, grayscale the scans and drop the pages you don’t need. Start at Compress PDF, and reach for grayscale, delete pages or split when the last few kilobytes are stubborn.
