About this tool
Shrink Large PDFs, Right in Your Browser
Drop a PDF and pick a preset: Smaller, Balanced, or Sharper. The compressor rasterises each page at the chosen resolution and re-encodes as JPEG inside a brand-new PDF. Best results on image-heavy scans and photo-rich PDFs, where 80–90% size reductions are common. Everything runs locally. Your PDF never leaves your device.
Why use OmegaPix
-
Huge savings on scans : A 60 MB scanned document typically drops to 5–10 MB on the Balanced preset with no perceptible quality loss for screen reading.
-
Pick your trade-off : Smaller (most aggressive), Balanced (default, looks great on screen), or Sharper (keeps detail at print zoom).
-
Nothing uploads : Other "compress PDF" sites send your file to a server. For tax returns, contracts, or scanned IDs, that's the real privacy risk. OmegaPix runs locally, your PDF stays on your device.
How it works
Drop your PDF
Single file or batch. Works on multi-hundred-page PDFs (give it a minute on large ones).
Pick a preset
Smaller / Balanced / Sharper. You can also fine-tune resolution and JPEG quality if you want.
Download the compressed PDF
Each result shows the size delta as a percentage. Filename gets a -compressed suffix so the original isn't overwritten.
Frequently asked questions
How does this compress so much?
It rasterises every page into a JPEG and rebuilds the PDF around those JPEGs. For image-heavy PDFs (scans, photo books, presentations) this gives massive savings because the original images get re-encoded at a controlled quality. For text-heavy PDFs the savings are smaller because the text was already small.
Will my PDF still be searchable?
No. The output is a sequence of images, so the text layer is gone. If you need a searchable compressed PDF, the right pipeline is to compress first, then run the output through an OCR tool to add a fresh text layer.
Which preset should I pick?
Balanced (default) is right for most uses, screen-reading sharp, sensibly small. Smaller if you need to hit an email cap and don't mind softer text. Sharper if the PDF is going to be printed or zoomed into for detail.
Will it work on text-only PDFs?
It works, but savings will be modest. Text-only PDFs are already small relative to scans. If the input is under a few MB, compression may not help much.
What about encrypted PDFs?
Password-protected PDFs need to be unlocked before they can be re-rendered. If you have the password, unlock first, then compress.
Are annotations, form fields, or signatures preserved?
No, rasterising flattens them all into the page image. Annotations become part of the picture (you can see them but not edit). Form fields lose interactivity. Digital signatures are invalidated. If you need to keep any of those interactive, this tool isn't the right one.
Is anything uploaded?
No. Compression runs locally via WebAssembly. Open DevTools → Network and you'll see no upload requests.
How big a PDF can I compress?
On desktop, ~200 MB or ~500 pages, whichever you hit first. Very large PDFs take a while, give them a minute or two.
When to use this tool
Email attachment limits
Many email systems cap attachments at 25 MB. A scanned 40-page contract often blows past that, compress it to a clean 5 MB and it sends fine.
Faster downloads
A web-published report at 50 MB is slow to load and costs you bandwidth. The same report at 6 MB downloads instantly without quality loss for screen viewing.
Storage cleanup
A folder of 200 scanned receipts can balloon to gigabytes. Run the batch through the compressor and reclaim most of it.
Mobile-friendly PDFs
Mobile data and storage are scarce, sending a colleague a 50 MB PDF over LTE is rude. A 5 MB compressed version is a much better citizen.
When not to use this tool
Text-only / vector PDFs
Small text PDFs are already compact, rasterising them makes them BIGGER, not smaller. OmegaPix detects this and returns the original instead, so you do not get a worse file.
PDFs you need to keep searchable
Compression converts pages to images, removing the text layer. If search/copy matters, keep the PDF uncompressed or run the output through OCR afterwards.
Interactive forms and signatures
Form fields, annotations, and digital signatures are all flattened into the page image. They become visible but lose functionality. Avoid compressing signed documents.
Technical details
The rasterise-and-rebuild approach
Each page is rendered via pdf.js to a canvas at the chosen DPI, then JPEG-encoded at the chosen quality, then embedded as a single image into a new PDF via pdf-lib. The new PDF contains only the rendered page images. No original text streams, no fonts, no vector data. This is how most browser-side PDF compressors work; it's mathematically optimal for image-heavy PDFs and provably bad for text-only PDFs.
Preset settings
Smaller: 72 DPI + JPEG q60. Most aggressive, ideal for email attachments where size trumps zoom-in quality. Balanced (default): 108 DPI + JPEG q75. Screen-sharp on phones and laptops. Sharper: 144 DPI + JPEG q85. Keeps detail when readers zoom in for inspection or pull pages for print.
When compression actually helps
Image-heavy PDFs are the killer use case. A 60 MB scanned document at Balanced typically drops to 5-10 MB (-85%). A 30 MB photo book to 4-6 MB. Mixed text-and-photo PDFs see 60-80% reductions. Pure text PDFs see no reduction. The tool detects this and keeps the original.
The keep-original safety net
After rebuilding, if the output is larger than the input, OmegaPix returns the original bytes with a "kept original" message. This prevents the absurd outcome of compression producing a bigger file. It also means small text PDFs are always safe to drop in, at worst, you get the source back.
OCR before vs after
If you need the output to remain searchable: compress first, then OCR the result with a tool like Tesseract or a paid OCR service. The OCR adds an invisible text layer back, giving you the compressed file size AND search. Doing it in the other order (OCR first, then compress) wastes the OCR work because rasterising destroys the text layer.
Your files stay on your device
PDF rendering (pdf.js), JPEG encoding, and PDF rebuilding (pdf-lib) all happen inside your browser. For financial statements, scanned IDs, or legal documents, this is the only safe way to compress, your file is never transmitted. Open DevTools → Network during the operation: zero outgoing requests carry PDF data.