How to compress a PDF without losing the useful parts
The right method depends on what the file contains. Text-heavy PDFs need a different treatment from scans and photo-heavy documents.
Free toolPDF CompressorCompare three real PDF sizes in your browser, keep text when you can, and save the one that fits without a watermark.Compress a PDFThe short answer
Start with a lossless rewrite. It can remove metadata and pack objects without changing selectable text. If the file is a scan, choose a stated image resolution and JPEG quality, then inspect the smallest type before saving.
Run the PDF Compressor when you want the candidates and their exact bytes in one place. It works in the browser, so the PDF itself does not go to a remote compressor.
What compression actually changes
| Method | Keeps | Can change | Good for |
|---|---|---|---|
| Lossless rewrite | Text, fonts, vectors and page structure | Metadata and object packing | Contracts, forms and reports where search matters |
| Page raster at moderate quality | Visible page appearance | Text selection, links and vector sharpness | Scans and image-heavy documents sent by email |
| Page raster at low quality | The broad page layout | Fine type, diagrams and image detail | A strict upload limit after visual inspection |
Three ways to do it
Use a browser tool
Browser tools are convenient when you need a result now and do not want to install software. Floi shows a Preserve candidate and two explicit raster candidates. The byte count appears before the save, and the PDF stays in the tab.
Use a desktop PDF app
Acrobat, Preview and other desktop apps can keep more document features, especially forms and links. Look for a reduced-size or optimised copy command. Keep the original because some apps overwrite the working file.
Use a command line
Ghostscript gives repeatable settings for a batch. Its presets trade resolution and image quality, and the output can lose features a desktop editor would preserve. Test one representative file before processing a folder.
What to check before you send the result
Search and copy
Try selecting a sentence. If it no longer selects, a raster mode changed the document into pictures.
Small type
Zoom a footnote and a table. A scan can look fine at fit-to-page and still lose the detail you need.
Links and forms
Open a link and type into a form field. Raster output cannot keep those interactions.
Why an already-small PDF can grow
A PDF stores text as a few drawing instructions and can reuse one font across every page. Turning that page into a JPEG stores a complete image instead. For a text-only document, that image may be larger than the original. The honest response is to keep the original or use a lossless rewrite.
What the launch test covered
On 17 October 2026 we ran the same synthetic three-page text PDF through iLovePDF, Smallpdf, CloudConvert and PDF24. Their output pages and controls are recorded in the launch research notes. The sample was public and contained no private material.
Floi's browser candidate sizes should be measured across text-heavy, scan-like and already-optimised PDFs before making an average reduction claim. The tool therefore reports the bytes for your own document instead of promising a universal percentage.
Frequently asked questions
Is PDF compression always lossy?
No. Removing metadata and rebuilding object streams can keep the document structure intact. Re-rendering pages into JPEG is lossy because text and vectors become pixels.
Why can compression make a PDF larger?
Rasterising a page that was mostly text can add more image bytes than the original text and font streams. A good compressor reports that result instead of hiding it.
What is the safest way to compress a scanned PDF?
Keep the original. Try a moderate raster setting, inspect small type and diagrams, and keep the original beside the saved copy. OCR text is a separate layer and may be lost if you rasterise after OCR.