Compress PDF
Drop PDFs, get smaller files. Text stays text.
Images inside the PDF are re-encoded as JPEG; text, fonts and vector graphics are not touched.
- 100% freeNo account, no daily quota, no paid tier.
- Nothing is uploadedFiles are processed on your device and never reach a server.
- No catch on the outputNo watermark, no forced downscaling, no size cap.
- Updated in the openEvery fix and every new tool is written down on the changelog page.
What this PDF compressor does
Almost all of a large PDF is images: the scan behind each page, the photos in a report, the screenshots in a manual. This tool opens the file, finds every image object, decodes it and re-encodes it as an optimised JPEG at the quality and size you choose — then writes it back in place. Text, fonts, vector drawings, links, bookmarks and page order are not touched, so the result is still searchable, selectable and prints the same. A 40 MB scanned contract typically comes out at 4–8 MB; a slide export with photos loses 50–80%. The number of images replaced is shown for each file, and if re-encoding an image would not make it at least 5% smaller, the original is kept.
Three presets cover the real cases. 'High quality' keeps every pixel and only re-compresses (JPEG quality 85) — for files that will be printed or archived. 'Recommended' limits images to 2000 pixels on the long side at quality 72, which is invisible on screen and right for email, portals and sharing. 'Smallest' goes to 1400 pixels at quality 55 for hard upload limits such as 2 MB or 5 MB. Everything runs in your browser with pdf-lib and the same MozJPEG encoder used by the image compressor; the file is never uploaded, there is no size cap and no watermark.
How to use it
Drop one or many PDFs onto the drop zone or click to browse; each file shows its page count and a first-page thumbnail. Choose a preset, click Compress, and watch the progress bar — a 300-page scan takes a minute or two because every page image is decoded and re-encoded on your machine. Results appear as a list with the before/after size and the number of images replaced. Download files one by one or all together as a ZIP.
Start with 'Recommended'. If a file is still too large, run it again with 'Smallest'; if you can see softness in photos, use 'High quality'. Compressing the same file twice with the same preset does nothing more — the tool will report it as already optimal. From the results you can go straight to merge, split or PDF to image without re-selecting the files. Your preset is remembered on this device.
Common pitfalls
A PDF that is mostly text, vector charts or embedded fonts has little to compress: the tool will report 'no compressible images' and the saving may be only a few percent from repacking the file structure. Images that carry transparency (soft masks), 16-bit, CMYK, JPEG 2000 or fax-encoded (CCITT) data are left as they are rather than risk changing how they render, so a file made of those will shrink less. Images smaller than 200 pixels — icons, logos, bullets — are skipped on purpose; re-encoding them saves nothing and blurs them. Files that require a password to open are skipped and reported; unlock them first. Compression is lossy for the images: keep the original if you may need to print at full resolution later.
Background: how images live inside a PDF
PDF was published by Adobe in 1993 and standardised as ISO 32000-1 in 2008. Inside a file, each image is an XObject stream with a dictionary describing its width, height, colour space and filter. Photographs are usually stored with the DCTDecode filter — a plain JPEG (ITU-T T.81, 1992) embedded byte for byte — while screenshots and graphics use FlateDecode, the same zlib/deflate compression as PNG (RFC 1950 and 1951, 1996). Scanners often embed pages at 300 or 600 dpi as JPEGs with generous quality settings, which is why a ten-page scan can be 30 MB. Re-encoding those images with a modern encoder at a lower quality and a sensible resolution, and writing the new stream back under the same object number, shrinks the file without altering anything that references the image.
Where compressed PDFs are needed
Email attachments, where 10–25 MB limits are common; government, university and job-application portals that cap uploads at 2, 5 or 10 MB; sending scanned contracts, ID documents and bank statements through messaging apps; keeping a project's PDF archive from filling a shared drive; and slide exports that ballooned because every photo went in at camera resolution. Because the file never leaves your device, the same tool is safe for tax forms, medical records and signed agreements — exactly the documents that upload services are worst suited to. If a PDF is a pure image scan and even 'Smallest' is not enough, convert the pages to images, compress them harder, and rebuild with image to PDF.
Frequently asked questions
Is my PDF uploaded to a server?
No. The file is opened, its images re-encoded and the PDF rewritten entirely inside your browser. It works offline once the page has loaded.
Will the text still be selectable and searchable?
Yes. Only image streams are replaced; text, fonts, vector graphics, links and bookmarks are copied unchanged.
How much smaller will my PDF get?
Scanned documents and photo-heavy files usually shrink 60–90%. Text-only PDFs shrink very little because there are no images to re-encode — the tool tells you when that is the case.
How do I compress a PDF to under 1 MB or 2 MB?
Use the 'Smallest' preset. If the result is still over the limit, the file is either very long or contains image types the tool leaves untouched; splitting it with the PDF splitter is the next step.
Why does it say the file is already optimal?
Re-encoding did not produce a smaller file — usually because the images were already compressed at a similar or lower quality. The original bytes are returned unchanged rather than a slightly worse copy.