What actually makes one compressor better than another
Almost every free tool can make a PDF smaller. The differences show up in the details: whether you can ask for a specific size, whether the result is still readable, whether your file is uploaded to a stranger's server, and whether a watermark or a signup wall appears at the download step. Those four things — control, quality floor, privacy, and friction — are what we scored.
Raw percentage reduction on its own is a misleading number. A tool that reports "82% smaller" on a scanned document may have simply rendered every page down to 72 DPI, which looks fine on a phone thumbnail and unreadable on paper. We treated any output where 10 pt body text became fuzzy at 100% zoom as a failure, no matter how small the file was.
How we tested
Each tool got the same three documents, in the same order, on the same connection:
- A 14 MB scanned contract — 12 pages photographed at high resolution. This is the file type with the most room to shrink.
- A 6 MB pitch deck — full-bleed photos, charts, and two embedded fonts.
- A 2 MB text-only report — no images at all, just vector text. The hardest file to compress, because there is almost nothing to throw away.
We recorded output size, time to result, whether the text stayed selectable, and what happened to the file after processing.
What we found
Scanned documents: 70–90% reductions are normal
The 14 MB contract came out between 1.4 MB and 4 MB across tools. Everything in that range is acceptable; the differentiator was legibility. Tools that let us name a target — "make this 2 MB" — landed close to it while keeping the scan crisp. Tools with only a "high compression" button overshot and produced a 900 KB file we could not read comfortably.
Image-heavy decks: expect 40–60%
Photos are already JPEG-compressed inside the PDF, so a second pass has less to work with. Recompressing them harder is exactly where visible artifacts appear — banding in gradients, halos around text on slides.
Text-only PDFs: don't expect miracles
The 2 MB report finished at about 1.7 MB almost everywhere. If a tool claims a huge reduction on a pure-text PDF, it has usually rasterised your pages into images, which makes the text unsearchable and often makes the file larger. If your text PDF is unexpectedly big, the cause is usually embedded fonts or an oversized page count, not image data — splitting it into sections is often the better fix.
The privacy question most comparisons skip
Compression is one of the few PDF operations that can run entirely inside your browser tab. When it does, the document never travels anywhere — there is no upload, no queue, and nothing to delete afterwards. PDF2Tool's Compress PDF tool works this way, which matters if the thing you are shrinking is a contract, a medical record, or a payslip.
Server-based compressors are not automatically unsafe, but you are trusting a retention policy you cannot verify. Read what happens to your file after processing before you upload anything you would not email to a stranger.
Common mistakes that waste your time
- Compressing twice. Running an already-compressed PDF through again mostly adds artifacts, not savings.
- Compressing before merging. If you are combining files, merge them first and compress the finished document once.
- Going below 120 DPI on scans. Text stops being reliably readable, and OCR accuracy collapses.
- Chasing a size limit you guessed. Check the real limit first — Gmail is 25 MB, most job portals are 5 MB, and many government forms are 2 MB.
Which one should you pick?
If you need a specific number — under 5 MB for an upload form, under 2 MB for a portal — use a compressor with exact-size targeting and let it iterate for you. If you just want a file that emails cleanly, a balanced preset is fine. If the document is sensitive, the browser-based option is the only one that removes the question entirely.
Frequently Asked Questions
How much can a PDF realistically be compressed?
Scanned documents commonly drop 70–90%. Image-heavy files drop 40–60%. Text-only PDFs rarely drop more than 15–20%, because there is very little redundant data to remove.
Does compressing a PDF reduce its quality?
It can. Compression works by lowering image resolution and re-encoding pictures. Vector text stays sharp, but scanned pages soften as DPI drops — which is why a readability floor around 120 DPI matters.
Is a free PDF compressor safe for confidential files?
It depends on where the work happens. A compressor that runs in your browser never transmits the document. One that uploads it depends entirely on the provider's retention and deletion policy.
Why did my PDF get bigger after compression?
Almost always because the tool rasterised text pages into images. If your PDF is text-based and already small, compression has nothing useful to remove.