PDF Compressor
Shrink a PDF by re-compressing its pages as images. Best for scans and image-heavy documents.
No data sent to serverRelated Tools
Read This First: What This Tool Does and Does Not Do
This compressor renders every page to an image and re-encodes it as JPEG. That approach is extremely effective on scanned documents and photo-heavy PDFs, where the bulk of the file is picture data that was stored at a higher resolution than anyone will ever look at.
It has a real cost: the output is a picture of your document, not the document. Text stops being selectable, copy-and-paste stops working, links stop being clickable, form fields disappear, and screen readers can no longer read the content. For a scanned contract you are archiving, none of that matters — it was already an image. For a text report you are sending to a colleague, it matters a lot.
If your PDF is mostly text and already small-ish, this tool will often make it larger, because a page of crisp text stored as vectors and fonts compresses far better than the same page stored as a photograph of itself. The result panel tells you the before and after size so you can check before downloading.
When This Is the Right Tool
- Scanned documents. Phone-scanned or flatbed-scanned pages are already images, so nothing is lost that was not already gone.
- Photo-heavy PDFs. Portfolios, catalogues, and property listings where the images dominate the file size.
- Presentation exports. Slide decks exported to PDF often embed full-resolution images nobody needs at print quality.
- Upload limits. When a portal caps attachments at 5 MB and the content matters more than the text layer.
And when it is the wrong tool: contracts that need selectable text, documents with fillable forms, anything that must remain accessible to screen readers, and files where someone downstream will search the text.
The Three Levels
- Light — renders at roughly 144 DPI with high JPEG quality. Still looks sharp on screen and prints acceptably. Modest size reduction.
- Balanced — around 108 DPI at moderate quality. The right default for sharing and archiving; text stays readable, photos stay presentable.
- Strong — around 80 DPI at low quality. Aggressive. Fine for reference copies and for getting under a stubborn upload limit, visibly soft if printed.
How to Use
- Pick a level. Start with Balanced.
- Drag in a PDF. It stays on your machine — nothing is uploaded.
- Click Compress PDF and watch the page counter.
- Check the before/after sizes. If the file got bigger, your PDF is text-based and this is the wrong approach.
- Preview the result in the embedded viewer, then download.
Technical notes
- Rendering: pdf.js draws each page to a canvas at the chosen scale
- Encoding: JPEG, with quality between 0.55 and 0.82 depending on level
- Page geometry: the original page dimensions in points are preserved, so paper size and orientation survive
- Transparency: flattened onto white, because JPEG has no alpha channel
- Not preserved: text layer, hyperlinks, bookmarks, form fields, attachments, and document metadata
- Encrypted PDFs: password-protected files cannot be opened and will report an error
FAQ
My compressed PDF got bigger. Why?
Your PDF is text-based. Text stored as fonts and vector instructions is astonishingly compact — far smaller than a photograph of the same page. Converting it to images works against you. Keep the original.
Can I still select and copy the text?
No. Every page becomes an image, so there is no text layer left. This is the fundamental trade-off of image-based compression, and it is why the tool warns about it up front.
Is there a way to compress without losing the text?
Yes, but not in a browser. Desktop tools like Ghostscript can downsample the embedded images while leaving the text layer intact. That is the better approach for mixed documents — it is simply not something browser PDF libraries can do today.
How much smaller will my file get?
For scanned documents, commonly 50–90%. For photo-heavy PDFs, often 60–80%. For text documents, usually negative. The result panel reports the actual figure before you download.
Does it work on password-protected PDFs?
No. Encrypted files cannot be opened for rendering, and you will see an error. Remove the password in your PDF reader first.
Is my document uploaded anywhere?
No. Rendering and re-encoding happen entirely in your browser. Confidential documents never leave your computer, which is the main reason to use a browser-based tool for this at all.
Why is a big PDF slow?
Every page is rendered and re-encoded one at a time on your CPU. A 100-page scan takes a while. Keep the tab in the foreground so the browser does not throttle it.
Can I compress just some pages?
Not selectively — every page is processed with the same setting. To treat pages differently, split the document first, compress the parts that need it, and merge the result.
Does compressing twice make the file smaller again?
Barely, and it damages the image further each time. If the first pass was not enough, choose a stronger level on the original rather than re-compressing the output.
⚠️ Reference Only
Output is generated based on your input and is provided for reference. Results may vary depending on your specific use case, edge cases, or environment-specific behavior. We do not guarantee accuracy of conversions, validations, or computed values.
Always verify critical outputs against official documentation or production environments. We are not responsible for any decisions or losses based on these tool results.