Many PDF compressors get their big savings by turning every page into a picture, so the text can no longer be searched, copied or read by software. Here is how to make a PDF smaller and keep its text, and how to check that you did.
You need a PDF under 2 MB for an email, or under 200 KB for an upload form. An online compressor gets it there, but afterwards Ctrl+F finds nothing, you cannot copy a sentence out of it, and the applicant tracking system that reads your CV sees a blank page.
That happens because the easiest way to shrink a PDF is to rasterise it: draw each page as a JPEG at a lower quality and wrap the images back into a PDF. The pages look almost the same, but the text is gone. Our own Compress PDF tool does exactly that in its Strong and Target size modes, and says so on the page, because for a scanned document it is the best option. For anything with real text, use one of the methods below instead.
First, find out why the file is big
Compression works on whatever is taking up the space, so it helps to know what that is.
| What is inside | Typical sign | What helps |
|---|---|---|
| Large photos or screenshots | A report or slide deck of several MB | Downsample and recompress the images |
| Scanned pages | Every page is one big image, often 300–600 DPI | Lower the resolution, then OCR if you need search |
| Full embedded fonts | A short text document of 1 MB or more, especially with Asian scripts | Font subsetting (Ghostscript does this) |
| Old revisions | A form or contract edited and saved many times | A clean rewrite drops them |
| Attachments | A paperclip icon or an Attachments panel | Remove them if the recipient does not need them |
A PDF that is pure text and already small, under about 300 KB for a few pages, has little left to gain. No method will shrink it dramatically without damaging it.
Method 1: export a smaller PDF from the original
If you still have the Word, PowerPoint or other source file, this is usually the best result for the least effort.
Microsoft Word and PowerPoint (Windows)
- Select a picture and choose Picture Format › Compress Pictures. Untick Apply only to this picture, tick Delete cropped areas of pictures, and choose 150 ppi (or 96 ppi for screen only).
- Go to File › Save As, choose PDF as the type, and under Optimize for choose Minimum size (publishing online).
Text stays text, and a document full of phone photos often shrinks by 80% or more. To stop Word saving huge images in the first place, set the default resolution under File › Options › Advanced › Image Size and Quality.
Google Docs, LibreOffice and others
Resize or compress large images in the document before exporting. In LibreOffice, File › Export as PDF has a JPEG compression quality setting and a Reduce image resolution option on the General tab.
Method 2: Ghostscript (free, any PDF)
Ghostscript is a free, open-source program that rewrites a PDF, recompressing its images and subsetting its fonts while keeping the text as text. It works on Windows, macOS and Linux. Once it is installed, run:
gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.6 -dPDFSETTINGS=/ebook \
-dNOPAUSE -dBATCH -dQUIET -sOutputFile=smaller.pdf original.pdfOn Windows the program is usually called gswin64c instead of gs. The -dPDFSETTINGS value sets the trade-off:
/screen: images at about 72 DPI. Smallest, fine on screen, poor in print./ebook: about 150 DPI. A good default for email and uploads./printerand/prepress: 300 DPI. Print quality, smaller savings.
This keeps the invisible OCR text layer of a searchable scan, which rasterising compressors throw away. Very occasionally a PDF with unusual fonts comes out with text that looks right but copies as gibberish, so always run the check at the end of this guide.
Method 3: desktop apps you may already have
- Adobe Acrobat Pro: the Compress PDF or Reduced Size PDF option keeps text. Optimized PDF gives you control over image resolution, font embedding and removing unused objects.
- Mac Preview: File › Export, then choose the Reduce File Size Quartz filter. Text survives, but the built-in filter compresses images very hard, so check that photos and small print are still readable.
Method 4: a lossless rewrite
If you cannot accept any change in image quality, a lossless rewrite can still help a little. It removes revisions left over from earlier saves, packs the file’s internal objects more efficiently, and drops hidden metadata. It does not touch the pages, so text, images and fonts are unchanged.
- In your browser: choose Keep text selectable (light) in our Compress PDF tool. The file is processed on your device and not uploaded.
- On the command line, with the free qpdf tool:
qpdf --object-streams=generate --recompress-flate --compression-level=9 original.pdf smaller.pdf
Savings are usually modest. The files that gain most are ones that were edited and saved many times.
Scanned documents: compress, then OCR
A scan with no OCR layer has no text to lose. It is a stack of photographs of paper. For these, rasterising compression is the right tool: our Compress PDF Strong or Target size mode often cuts a large scan to a fraction of its size while keeping it readable.
If you need the result to be searchable, run OCR afterwards. The free OCRmyPDF adds a text layer and can optimise at the same time:
ocrmypdf --optimize 3 compressed-scan.pdf searchable.pdfDoing it in that order keeps the OCR layer intact, because nothing rasterises the file after the text has been added.
Check the text survived
Do this on the compressed file before you send it. It takes ten seconds.
- Open it and press Ctrl+F (Cmd+F on a Mac). Search for a word you can see on the last page.
- Try to select and copy a sentence, then paste it somewhere. It should come out as the same words, not boxes or gibberish.
- For a longer document, drop it into our PDF to Text tool. It extracts the text from every page and flags any page that has none, which shows straight away if some pages were turned into images.
Hitting a hard size limit? If an upload form needs a CV under 200 KB and no text-preserving method gets there, check whether the form also accepts DOCX, or whether the photo on your CV is the problem. A headshot pasted at full camera resolution can be most of the file.
Frequently asked questions
Why can’t I search my PDF after compressing it?
The compressor probably rasterised it: it drew each page as an image and threw away the text. The PDF looks the same but contains only pictures. Go back to the original and use a method that recompresses the images and keeps the text, such as Ghostscript, Acrobat’s Reduced Size PDF or re-exporting from Word.
How small can a text-only PDF get?
A PDF that is mostly text is already small, often 50 to 300 KB for a dozen pages. Most of that is embedded fonts. If it is much bigger than that, there are images or attachments inside, and those are what to compress.
Does compressing a PDF affect a digital signature?
Yes. Any rewrite invalidates an existing digital signature. Compress first, then sign.
Can a compressed scan be made searchable again?
Yes, with OCR (optical character recognition), which adds an invisible text layer over the page images. OCRmyPDF does this for free, and Acrobat Pro has it built in.