How to Compress a PDF Without Losing Quality (We Tested 3 Real PDFs)
"Compress a PDF without losing quality" sounds like one job. It's actually two, and which one you need depends on what's inside your PDF. A 20-page report typed in Word and a 20-page document scanned on a phone can be the same file size and still need completely different treatment.
To show the difference with real numbers instead of promises, we ran three very different PDFs through Compressnfly's PDF compressor and recorded exactly what happened.
The short answer
- A PDF made from typed text is already small. Lossless clean-up trims a bit more and nothing changes visually.
- A scanned PDF is really a stack of photos. Lossless clean-up does almost nothing; shrinking it means re-rendering the pages at a lower resolution.
- For on-screen reading, about 40–60 KB per scanned page stays clearly readable. Below about 25 KB per page, text gets soft.
What actually makes a PDF big
Text is cheap. Our 19-page text-only handbook, exported from Word, was just 100 KB. That's roughly 5 KB per page, because text is stored as characters plus a font, not as pictures.
Size comes from images, and there are three common sources:
- Scanned pages. A scanner or scanning app saves every page as a photo, often at 300 DPI. Our 8-page test contract, built the way a scanning app saves pages (each page a 300 DPI color JPG), weighed 9.49 MB, more than a megabyte per page, even though it's just dark text on white paper.
- Photos pasted into documents. Drop phone photos into a Word report and export it, and the photos come along at close to full resolution. Our 9-page inspection report with six photos (three phone pictures, each used twice) was 2.31 MB as a PDF.
- Embedded fonts and leftovers. Fonts, form data and old edit history add some weight, but in our tests they were never the main problem.
Our test: three real PDFs
We compressed each file twice: once with a lossless clean-up (text stays text), and once with the Small Size setting, which re-renders each page as an image.
| Original | Lossless clean-up | Small Size (pages as images) | |
|---|---|---|---|
| Text-only handbook, 19 pages | 100 KB | 64 KB (−36%) | 64 KB: re-rendering would have made it bigger, so it was skipped |
| Word report with 6 phone photos, 9 pages | 2.31 MB | 2.30 MB (−0.4%) | 625 KB (−74%) |
| Scanned contract, 8 pages | 9.49 MB | 9.48 MB (−0%) | 804 KB (−92%) |
That table is the whole story. Lossless clean-up is the "no quality loss" option, and it works well on text-heavy files. On anything built from images, it barely moves the needle, because the images themselves are the weight.
The two kinds of PDF compression
1. Lossless clean-up: zero quality loss
The PDF is rebuilt with its internal objects packed more efficiently. Every word stays selectable and searchable, links keep working, and nothing looks any different. It's the safe default, and it's what Compressnfly runs first in every mode. The catch: it can only remove waste. If a file has no waste, you'll see little change, like our scanned contract, which dropped by about 1.5 KB.
2. Re-rendering pages: big savings, small trade-off
Each page is drawn as an image at a lower resolution and saved as a compressed JPG inside a new PDF. This is how a 9.49 MB scan becomes 804 KB. The trade-off is that the text becomes part of a picture: you can still read it, but you can no longer select, copy or search it. Compressnfly tells you when this happens, with a note under the result.
For a scan, that trade-off costs you nothing, since the text in a scan was never selectable in the first place. For a typed document, think twice before using it.
How small can a scanned PDF get and stay readable?
Upload portals love limits like "PDF under 200 KB" or "under 500 KB". So we pushed the same 8-page scanned contract to exact sizes and looked at the pages afterward:
| Target | Result | Per page | How it looked |
|---|---|---|---|
| 2 MB | 1.31 MB | ~168 KB | Crisp; fine for printing |
| 1 MB | 894 KB | ~112 KB | Sharp on screen |
| 500 KB | 346 KB | ~43 KB | Clearly readable on screen, slight softness when zoomed in |
| 200 KB | 167 KB | ~21 KB | Readable, but visibly fuzzy; small print takes effort |
Notice the tool doesn't stop exactly at the limit. When it has room to spare, it keeps the sharper version that still fits instead of degrading the file just to land on the number.
The rule of thumb we took from this: divide your limit by the number of pages. At 40 KB or more per page, a typical text scan stays comfortable to read. At 20 KB per page it's still legible but rough. If your 10-page scan has to fit under 200 KB, consider whether every page is needed. Sending 5 key pages at 40 KB each reads far better than 10 pages at 20 KB.
You can do this yourself on the compress PDF to 200 KB, 500 KB or 1 MB pages, which have the target size already set.
How to compress a PDF on Compressnfly
- Open the PDF compressor and drop in your file. It's read by your browser, not uploaded anywhere.
- Pick how you want to compress:
- By quality level: High Quality, Web Optimized and Balanced do the lossless clean-up only, so text stays selectable. Small Size also re-renders pages as images when that saves at least 20%.
- To an exact size: pick a limit like 200 KB or 1 MB. The tool tries the lossless clean-up first and only re-renders pages if that isn't enough to fit.
- Check the before/after size and any note under the result, then download.
Open the compressed file and zoom in on the smallest text before you send it anywhere important. It takes ten seconds and saves an awkward resend.
Make smaller PDFs from the start
Prevention beats compression, especially for scans:
- Scan text documents at 150–200 DPI, in grayscale. 300 DPI color is for photos and artwork. For a typed letter it mostly records the texture of the paper.
- Use your scanning app's "document" mode. It boosts contrast and whitens the background, which also compresses better.
- Shrink photos before you paste them into Word. A report doesn't need 4032-pixel-wide images. Our image compression test shows what each size looks like.
- Export, don't "print to PDF", from Word. Export usually produces cleaner, smaller files, and it keeps links working.
When not to compress a PDF
- Fillable forms. Re-rendering pages turns form fields into pictures. Stick to the lossless levels, or fill the form before compressing.
- Documents people need to search or copy from: contracts under review, reports with data, anything that will be quoted. Keep the text layer.
- Files headed to a professional printer. Print shops want full resolution. Send the original.
- Password-protected PDFs. The pages can't be rendered without the password, so remove the protection first (you need to know the password to do that).
FAQ
Will compressing my PDF make the text blurry?
Not with the lossless levels: text stays exactly as it was. With Small Size or a tight size target, pages become images, so text sharpness depends on how much room each page gets. Use the per-page numbers above as a guide.
Why did my PDF barely get smaller?
Either it was already efficient (common for text PDFs), or it's made of images and you used a lossless level. Try Small Size or set an exact target.
Is it safe to compress confidential PDFs?
On Compressnfly, the file is processed inside your browser and never sent to a server. The page downloads the compression code; your document stays on your device.
Can I get the text back after re-rendering?
Not from the compressed copy. Keep your original. If you need searchable text from a scan, that requires OCR software, which is a separate step from compression.
Related: compressing a PDF to fit an email limit and getting documents under upload limits.