How to Compress a PDF for Email Without Losing Quality

The reason your 38 MB PDF won't go through is not actually that it's too big — it's that nine-tenths of those bytes are an embedded scan of a single contract you photographed at 12 megapixels with the iPhone camera. Strip that one image down from 4032×3024 to a sane 2000×1500 and re-encode it at JPEG quality 75, and you're under 6 MB without touching the text. Most "compress my PDF" failures come from squeezing the wrong thing — running a generic compressor that downsamples every embedded image including the company logo to mush, instead of identifying what actually needs to shrink.

This guide covers the email-attachment limits that matter for the four big providers, the math behind why a PDF is the size it is, what to compress first (and what to leave alone), and four worked examples from real document types. Run yours through ScoutMyTool's client-side PDF compressor once you know which knob to turn — the file never uploads, so even sensitive attachments can be processed without touching a third-party server.

The Email Attachment Limits You Actually Hit

Every major email provider enforces a per-message attachment cap that is much smaller than the underlying mail-server limits would suggest. The numbers below are the practical caps as of 2026 and the source-of-truth links to confirm:

• Gmail: 25 MB outgoing, 50 MB incoming (per Google Workspace attachment-size docs). Above 25 MB Gmail offers Drive integration that uploads the file and inserts a share link, but recipients on other providers may not be able to access it.
• Outlook (Microsoft 365): 20 MB default for personal accounts, configurable up to 150 MB for enterprise tenants per the Microsoft attachment-size guide. Many corporate inbound filters reject anything over 35 MB regardless of the sender's tenant settings.
• iCloud Mail: 20 MB default, with Mail Drop offering up to 5 GB by uploading to iCloud and emailing the recipient a 30-day download link.
• Yahoo Mail: 25 MB, with no integrated large-file fallback for free accounts.

The smallest common denominator is the 20 MB Outlook cap, and the safe target for anyone-to-anyone sending is under 18 MB to leave headroom for MIME encoding overhead (base64 inflates the actual byte count by roughly 33%). If your recipient is on Gmail, you can squeak by at 24 MB; if on Outlook corporate, you may need 15 MB or less to clear inbound filters.

Why Your PDF Is the Size It Is

Open any PDF and ask "where do the bytes live?" Almost every oversized PDF answers the same way: 80–95% of the file is image content. The text and vector layer of even a hundred-page document is typically under 200 KB. The fonts, if subset properly, add another 100–500 KB. Everything beyond about 1 MB is images.

PDF images come in three buckets, and the right compression strategy depends on which bucket dominates yours.

• Photographic images (camera scans, screenshots, embedded photos) are typically encoded as DCT (i.e., JPEG) inside the PDF. They re-encode well at lower quality, so this is the easiest thing to shrink with the highest visible-quality return.
• Bitonal images (black-and-white scanned text, fax-style documents) are usually encoded with CCITT Fax G4 or JBIG2 lossy compression. They're already small per page; downsampling rarely buys much, and lossy JBIG2 carries a now-infamous risk of swapping visually similar digits ("6" → "8") in numeric tables.
• Vector graphics (diagrams, line art, native PDF text) take effectively zero space relative to images. Don't waste time trying to shrink them.

A 2-page PDF that is 35 MB is almost certainly two phone-camera scans at 12 MP each. Reducing them to 200 DPI at the original page size brings the file under 5 MB without any visible degradation at normal viewing distance. A 200-page text-only PDF that is 1.2 MB is already as small as it's going to get without losing the embedded fonts.

Step-by-Step Using ScoutMyTool

The PDF compressor on ScoutMyTool offers three preset levels: light (downsample images to 300 DPI, JPEG quality 90), medium (200 DPI, quality 75), and aggressive (150 DPI, quality 60). Drop the file in, pick a level, and download the result — everything happens in your browser.

For mixed documents (cover photo + text body), medium is almost always the right call. Light does too little; aggressive starts producing visible JPEG ringing on screen. For pure-text PDFs, none of the levels will do anything useful — text PDFs aren't image-heavy, and there's nothing to compress. For pure-photo bundles (camera scans of receipts, signed contracts, ID copies), aggressive compression typically still leaves the document fully legible and can shrink a 30 MB scan bundle to under 5 MB.

If a single page is the problem, split the PDF first, compress just the offending page, and then re-merge. This preserves the rest of the document at full fidelity.

Worked Examples

Example 1 — A 38 MB phone-scanned contract. A real estate agent has a 6-page contract scanned with the iPhone Notes app at native camera resolution. Each page is roughly 6 MB. Goal: under 25 MB for Gmail. Action: medium compression downsamples each page from ~3024×4032 to ~1700×2200, with JPEG quality 75. Result: 5.8 MB total, every signature and clause still readable on a 13" laptop screen at 100% zoom.

Example 2 — A 12 MB sales deck with embedded images. A B2B account exec is sending a quarterly business review deck to a customer on Outlook. The deck is 22 slides, mostly text with eight embedded product photos. Goal: under 18 MB for safe Outlook delivery. The deck is already under, so do nothing. (The most common compression mistake is over-compressing files that don't need it; the second-most-common mistake is degrading them to "save bytes" the recipient never asked for.)

Example 3 — A 47 MB accordion of receipts. A consultant submits a year-end expense report as a single PDF: 84 receipt scans concatenated, ~600 KB each. Goal: under 25 MB for Gmail. Action: aggressive compression (150 DPI, quality 60) downsamples each receipt to a still-legible 800×1200 JPEG. Result: 14 MB. Every dollar amount, vendor name, and date is still clearly readable. Tax-audit acceptability: yes — the IRS does not require any specific resolution for digital receipt copies, only that the content be legible.

Example 4 — A 31 MB legal exhibit bundle that should not be lossy-compressed. A paralegal has a 200-page litigation exhibit bundle that includes original signed pages with handwritten notations. Goal: under 25 MB. Action: do not use lossy JBIG2 compression on the text-bearing pages — the substitution-error risk on numbers in legal evidence is unacceptable. Instead, downsample only the embedded color exhibits, leave the bitonal text pages alone, and re-merge. Result: 22 MB, with original page integrity preserved on every text page.

Common Pitfalls

The biggest pitfall is using "compress" as a synonym for "make smaller at any cost." Always check what's actually in the file before compressing. If the file is dominated by text, no compressor on earth is going to shrink it meaningfully — and most will quietly degrade the embedded fonts in pursuit of bytes that aren't there.

The second pitfall is double-compression. Re-compressing a JPEG that's already inside a PDF at quality 60 produces visible ringing artifacts that compound across passes. If you've already compressed once, don't compress again — restart from the original document.

The third is using lossy JBIG2 on text-bearing scans. The Xerox-scanner numerical-substitution scandal of 2013 was a real-world case of JBIG2 substitution errors silently changing "6" to "8" on architectural blueprints. Most modern compressors avoid lossy JBIG2 by default for exactly this reason, but a few still offer it as an option. Don't use it on legal, medical, financial, or engineering documents.

The fourth is uploading sensitive files to cloud-based compressors. The whole point of compression is usually to send a confidential document; uploading it to a third-party server first negates the privacy gain entirely. A client-side compressor running in your browser is the only path that doesn't introduce a new third-party copy.

Frequently Asked Questions

Q: Why is my PDF still huge after running it through a compressor? A: Either it's already optimized (text-only PDFs can't get much smaller), or the compressor is set to a "light" preset that doesn't actually downsample images. Try the medium or aggressive preset, and confirm by checking the file size of the dominant image — if a single embedded image is 5 MB and you're not touching image quality, you're not going to see meaningful savings.

Q: What's the maximum attachment size in Gmail? A: 25 MB outgoing, 50 MB incoming. Above 25 MB Gmail switches to Google Drive integration, which inserts a share link instead of an attachment. Recipients on Outlook or iCloud can usually open Drive links if they have a Google account but may not if their organization blocks Drive at the network level.

Q: Does compressing a PDF reduce text quality? A: No, if the compressor only touches embedded images. Text in PDFs is stored as font references and character positions, which are already maximally compressed. A reputable compressor leaves text alone and operates only on the image stream. If your text looks blurry after compression, the compressor was buggy and rasterized the page.

Q: Can I compress a PDF without uploading it to a server? A: Yes. Browser-based PDF compressors using PDF.js and the WebAssembly port of pdfium can re-compress images entirely client-side. This is the only privacy-safe option for HIPAA-covered, attorney-client-privileged, or financial documents.

Q: What DPI should I use for compressed scanned receipts? A: 200 DPI is the sweet spot for receipt-style documents — readable at 100% zoom on a laptop, scannable by most OCR tools, and a fraction of the file size of a 300 DPI scan. The IRS sets no specific DPI floor; the requirement is legibility, which 200 DPI satisfies for any standard-size receipt.

Q: My PDF has both text and photos. What compression level is safe? A: Medium (200 DPI, JPEG quality 75) is a safe default. It downsamples photos to a level that's still sharp at 100% screen zoom, leaves text untouched, and typically achieves 70–85% size reduction on phone-scanned documents. Light is too cautious; aggressive risks visible artifacts on photos viewed at full zoom.

Q: Can I compress a PDF with embedded form fields? A: Yes, and the form-field structure (which is metadata, not image content) survives compression untouched as long as the compressor preserves the document structure. Avoid compressors that "flatten" the PDF as part of their pipeline — that destroys interactivity. ScoutMyTool's compressor preserves form fields, annotations, and bookmarks by default.

Wrapping Up

Compressing a PDF for email isn't a one-knob operation — it's a three-step diagnosis: figure out which provider you're sending to (and the resulting MB cap), figure out where the bytes are actually living in your file (almost always images), and pick the compression level that shrinks those images without breaking what matters. The client-side PDF compressor handles the technical side; this guide handles the rest. Once you know what to compress and why, getting under any provider's attachment cap is rarely more than one click of work.