Reduce PDF file size — when your PDF is too large to send

The short version

1. Know your target size. Gmail 25 MB, Outlook 20 MB, most web forms 10 MB, WhatsApp 100 MB. The table below covers the rest.
2. Look at the PDF. Scan through it. Lots of photos or scanned pages? "Image-heavy." Mostly text and maybe a logo? "Text-heavy." Different paths.
3. Image-heavy → deep-compress. Open the PDF compressor, tick Re-encode images, leave quality at 85. 60-90% smaller is normal.
4. Text-heavy → wrapper-compress. Same tool, default settings. Expect 5-20% smaller. If it was a fresh Word export, maybe 0% — and we'll tell you.
5. Still too big? Split the PDF and send in two emails, or host it and send a link (Google Drive, Dropbox, WeTransfer).

The real size limits, by where it's going

Every "reduce PDF size" guide starts with the same sentence: "Email has a 25 MB limit." That's only true for Gmail, and only when both ends are Gmail. Here's the real table — check the column that matches where your PDF is going before you start compressing:

The number you actually need is usually 10 MB or less (forms and portals), not 25 MB (email). If you're emailing between two Gmail accounts, 25 MB really is the ceiling — over that, Gmail quietly rewrites the attachment as a Drive link, which some corporate firewalls strip.

Diagnose before you compress

Every PDF is one of two shapes, and which one determines whether any tool on Earth can help you:

Image-heavy PDFs

Every page is, under the hood, a photograph. Typical sources:

• Phone-camera scans of forms, receipts, contracts
• Flatbed-scanner exports (the office scanner, doctor's office, bank)
• "Scan to PDF" from a multifunction printer
• PowerPoint with embedded photos exported as PDF
• A textbook chapter, journal article, or court record scanned in bulk

Tell: try to select text inside the PDF. If the text doesn't highlight — if your cursor draws a box over letters like they're a picture — it's image-heavy. Also tells: the file is 5 MB+ and pages are visually grainy at 100% zoom.

What works: deep-compress. Re-encode each page as a lower-quality JPEG. Realistic savings: 60-90%. A 22 MB scan of a 10-page contract lands around 3-5 MB. Our PDF compressor does this when you tick Re-encode images in settings.

Text-heavy PDFs

The text is actual text, not pixels. Typical sources:

• Microsoft Word export (Save As PDF)
• Google Docs / Pages / LibreOffice export
• LaTeX (academic papers, technical docs)
• Software-generated invoices, reports, contracts
• Ebooks (EPUB-to-PDF)

Tell: you can highlight, copy, and paste the text. Files are typically 500 KB - 5 MB.

What works: wrapper-compress. Modern exporters already use efficient object streams, so realistic savings are 5-20%. A 2 MB report goes to about 1.7 MB. That's the honest number. If a tool promises 60% off a Word-exported PDF, it's either rasterizing your text (destroying selection) or lying. Ours won't do either.

The deep-compress workflow (image-heavy PDFs)

This is the path people asking "why won't my PDF shrink?" almost always need.

1. Open the PDF compressor.
2. Drop the PDF.
3. Expand Advanced settings.
4. Tick Re-encode images (this is the deep-compress switch).
5. Leave Quality at 85. It's the sweet spot — visually indistinguishable from the source on phone and laptop screens, and gives you most of the savings.
6. Click Compress. Download.

Typical outcomes on image-heavy source material:

• Phone scan of a 5-page contract, 12 MB → ~1.5 MB (88% off)
• Multi-page scanned tax return, 25 MB → ~3-5 MB (80-88% off)
• PowerPoint deck with 30 photo slides, 18 MB → ~5 MB (72% off)
• Architecture portfolio with renderings, 60 MB → ~12 MB (80% off)
• Medical imaging report with X-ray thumbnails, 30 MB → ~6 MB (80% off)

If q=85 overshoots your target (say you need under 2 MB for a visa form and you're at 2.4), drop quality to 75. Below 65 you start seeing visible softening on dense text; do that only when the size budget is non-negotiable and the document is for quick review, not print.

Important: deep-compress turns your vector text into a bitmap. Selection and search stop working. If you need the file to stay searchable, either keep the original for your records and send the compressed copy, or OCR the output afterwards (that's on our roadmap; today you can pair our output with Acrobat's "Recognize Text" or command-line Tesseract).

When compression hits a wall

Sometimes you'll run our tool, get the output, and the file is nearly the same size. When that happens you're probably looking at one of these:

• The PDF is already optimized. Modern exporters (Word 2013+, Pages, Google Docs, LaTeX with pdftex) write compact PDFs. Wrapper-compress has nothing left to strip. A 600-KB report won't drop to 400 KB no matter what tool you try.
• The PDF is all text with no images. No JPEGs to re-encode means deep-compress can't help either.
• The images inside are already JPEGs at low quality. Re-encoding an already-lossy JPEG saves a little but introduces more artefacts; at some point you're just degrading quality for 3% off the file size.
• Embedded fonts you can't drop. Fonts used for technical documents (LaTeX, legal) sometimes pull in 200 KB of glyph data that legitimately has to be there.

Our compressor is explicit about this: if wrapper-compress can't save at least 1%, we return the original file with this notice:

“This PDF is already well-compressed — we returned the original. Try the deep-compress option (re-encode images) for image-heavy PDFs.”

That saves you from the common scam on free PDF sites: they rename your file yourfile-compressed.pdf, hand it back at the exact same size, and hope you don't check with ls -l or File Explorer's size column. If your file's at this wall, the answer isn't another compressor — it's the alternatives below.

When compression isn't enough — alternative paths

If you've deep-compressed and you're still over the limit, compression alone won't save you. The fixes:

Split the PDF

A 40 MB appendix packet you're emailing as one attachment can go as two 20 MB emails — part 1 and part 2 — and no recipient will mind. Our PDF splitter does range-based or per-page splits; pick roughly-even splits below your target limit. Works well for:

• Contract bundles (split at section breaks)
• Scanned receipt archives for expense reports
• Photo-heavy portfolios — split by project
• Court filings where the receiving system caps each submission

Host the file, send a link

If you don't control the recipient's inbox limit (corporate Outlook, government form), hosting beats compressing. Quick options:

• Google Drive — free 15 GB; shareable link; Gmail does this automatically above 25 MB.
• Dropbox — free 2 GB; Transfer feature (free 100 GB per transfer, 7-day expiry) works without signup.
• WeTransfer — free up to 2 GB per transfer, 7-day expiry, no account. Fastest path for a one-off send.
• Smash — free unlimited size, 14-day expiry, shows a preview before download. Good for design files.
• Microsoft OneDrive — free 5 GB; native if both sides use Outlook.

Most people default to emailing the attachment. For anything over 15 MB that's going to a mixed-client audience, a hosted link is usually the better move — no corporate filter strips it, no quota gets blown on the recipient's side.

Merge first, compress second

If you have multiple source PDFs and you want one compressed output, merge them first then deep-compress. The deep-compress pass uses shared compression tables more efficiently across pages, producing a smaller file than compressing each PDF separately and then merging.

Batch workflow for repeat senders

If you're doing this once a week (accounting, legal ops, real estate, medical coding), the cycle is:

1. Scan or receive the PDFs into a local folder.
2. Drop the whole folder on the compressor — multi-file is on every plan.
3. Leave settings at "Re-encode images, quality 85" as the preset if you're mostly handling scans.
4. Download the ZIP.
5. Rename if needed and send.

Paid plans let you save your own compression preset (e.g. "invoices: wrapper only" vs "scans: re-encode q75") so you aren't re-ticking settings every session. Pricing details on the pricing page.

Honest compare — how we stack up

Tools in this category cluster at three levels: wrapper-only (too many free sites), wrapper + deep-compress done honestly (us, Acrobat, iLovePDF Pro), and CLI (ghostscript — fastest, but command-line).

Honest pick: Acrobat Pro if you're in the Adobe ecosystem already. Ghostscript if you're a CLI person with a batch of 100+. For everyone else — single-shot or weekly workflow, no subscription, a tool that tells you the truth when it can't help — our PDF compressor is the shortest path.

Works well / doesn't work

Works well

• Scanned documents, receipts, contracts heading to email or forms
• Phone-camera PDFs ("Scan Documents" in iOS Notes / Android)
• PowerPoint or Keynote exports with embedded photos
• Architecture, design, and photography portfolios
• Batch workflows for repeat senders (legal, accounting, medical)

Doesn't work well yet

• Keeping text searchable after deep-compress (OCR step is roadmap)
• Form-field preservation under aggressive deep-compress — fillable fields can flatten
• Encrypted / password-protected PDFs (remove password first with Acrobat)
• PDFs over 1.5 GB on free tier — see pricing

Common questions

Why is my PDF 20 MB when it's only 5 pages?

Almost always: it's a scan, not text. Every page is a high-resolution photograph, typically 2-5 MB each. Deep-compress at q=85 usually takes it to 1-1.5 MB total. If you can't highlight text in the PDF, that's your situation.

Will compressing the PDF affect print quality?

Wrapper-compress, no — it's lossless. Deep-compress at q=85 is fine for office printing up to Letter/A4; at q=75 you may see slight softening on halftones or thin text; below q=65, print starts to look compressed. For archival or professional print output, keep the original and send the compressed copy separately.

Can I compress a password-protected PDF?

Not directly. Most compressors, including ours, need to read the content streams — encryption blocks that. Remove the password first (Acrobat: File → Properties → Security → No Security; or via a dedicated unlock tool with your password), compress, then re-add the password if needed.

What's the difference between reducing size and optimizing?

"Optimize" usually means the lossless wrapper path — linearize, drop unused objects, strip metadata. Gains 5-20%. "Reduce size" colloquially includes the deep-compress / re-encode path, where images are re-sampled and re-encoded lossily. Gains 60-90% on image-heavy PDFs. Our tool labels them clearly as separate toggles.

Is it safe to upload a confidential PDF?

Our processing runs server-side with auto-delete inside the hour — no disk writes beyond the processing window, no database storage, no analytics on file contents. Details in our privacy policy. For work under NDA or legal hold where no third-party server is acceptable, ghostscript CLI on your own machine is the safer default — same algorithm, no network hop.

Will this break a digital signature on the PDF?

Yes — any re-save (wrapper OR deep-compress) invalidates the cryptographic signature on the PDF, because the signature hashes the byte contents. This is true of every compressor, not just ours. If the signature matters (tax filings, signed contracts), compress a copy for sending and keep the signed original.

What if the deep-compress output is still too big?

Drop quality to 75, then 65. If still too big, split the PDF into two files. If still too big, host it and send a link. Past that, the honest answer is that the content — particularly if it's photo-grade images — legitimately needs the bytes.

Ready?

Reduce PDF size →. Drop the file, tick Re-encode images if it's a scan, click Compress. Free tier works for files up to 4 MB input. No signup. If we can't save at least 1% honestly, we'll hand you back the original and point you at the next step.