PDF

If you want a single file out, use Extract Pages or Delete Pages. Pick whichever list is shorter: extract when you only need a few pages, delete when you only want to drop a few. Neither modifies your original — both write out a new PDF.

If you want several files out, use Split PDF. It can split by page range, every N pages, or one page per file, and hands you everything back as a single ZIP. Merge PDF is the opposite direction: drop in several PDFs and drag the thumbnails to set the order.

To rearrange the page order or fix the orientation of individual pages, use Organize PDF — drag-and-drop reordering and per-page rotation live on the same screen.

When an email or a submission form rejects your file, use Compress PDF. It has three strengths, and they work in fundamentally different ways. The lightest one only strips document info, XMP metadata, and embedded page thumbnails, so the appearance and the selectable, searchable text are preserved exactly. For text-heavy PDFs, always try this first.

The medium and strong settings re-render every page as an image and rebuild the PDF from JPEGs. For scan-heavy documents this shrinks the file dramatically, but because each page becomes a picture, the text is no longer selectable, searchable, or copyable. Choose according to what the file is for — a preview lets you judge the quality before you commit.

Need to compress several PDFs at once? Drop them all in and you get a single ZIP back.

PDFs come in two flavours: ones with real embedded text, and ones that are just scanned images of paper. They need different tools. Start with PDF Text Extractor — if text comes out, it was embedded, and you can copy or save it directly.

If nothing comes out, you have a scanned PDF. Use OCR Text Recognition, which reads characters out of the page image with a detection model (DBNet) and a recognition model (CRNN). The recognition model runs against an English dictionary, so English text only — Japanese, Chinese, and other scripts are not supported. The models run on WebAssembly, so Chrome, Edge, Safari, and Firefox all work.

PDF to Image writes out every page as a PNG. You can pick 72, 96, 150, or 300 DPI. For slide thumbnails or blog embeds, 96–150 DPI is usually right; for print or for images people will zoom into, go with 300. Higher DPI means a sharper image but also a bigger file and a longer wait.

Going the other way, Image to PDF turns photos of documents or a pile of screenshots into one file. Line up multiple PNG or JPEG images in the order you want and get back a single PDF — handy for turning phone snapshots of paperwork into something submittable.

PDF Markup lets you drop blackout rectangles onto a page and paste in images or seal stamps — including stamps you made in the Seal Stamp Maker, pasted straight in with Ctrl+V.

One caveat matters a great deal here. The blackout draws a filled black rectangle on top of the page. It hides the content visually, but the text underneath still exists inside the PDF and can be recovered with a text extractor or a plain copy-paste. To actually destroy the underlying text, re-flatten the marked-up file: run it through Compress PDF on medium or strong, or send it through PDF to Image and back with Image to PDF. Once the page is a picture, the text data is gone.

To discourage redistribution of a document you are handing out, use PDF Watermark. The text is drawn in the browser before being embedded, so non-Latin scripts such as Japanese work as-is, and you can tune opacity, size, colour, angle, and placement.

PDF passwords come in two kinds: a user password, needed to open the file at all, and an owner password, which only restricts actions like printing or copying. Unlock PDF handles both, but not identically. Removing owner-password restrictions leaves the text and vector content intact. A file encrypted with a user password requires you to enter the correct password, and after decryption each page is rebuilt as an image — so text in the output is no longer selectable or searchable.

If you do not know the password, the file cannot be opened. There is no password cracking or guessing here.

Another common annoyance: a PDF from a scanner or a phone app looks fine in your viewer but shows up sideways in other tools. That happens when only the page's rotation flag is set while the content itself was never rotated. Normalize PDF redraws the pages in their visual orientation, which also makes them behave properly in the markup tool.

The PDF structure is read and written with pdf-lib; page rendering, text extraction, and decryption of encrypted PDFs go through PDF.js. Both run inside the browser, and the file you choose is simply loaded into memory as a JavaScript ArrayBuffer. There is no upload path in the code at all.

You do not have to take our word for it. Open your browser's developer tools, keep the Network tab visible, and process a file. Apart from the resources the page itself needs, you will not see a request carrying your document anywhere.

Heavy work is pushed into Web Workers so the interface stays responsive on large files. For AI-backed features such as OCR, the model files themselves are downloaded to your browser and inference runs on your device. Those models are served from our own origin — no third-party CDN and no external API is involved.

OFFiler is a PWA. Once you have visited it, the tools keep working in airplane mode or on a flaky connection. Add it to your home screen or desktop and it launches like an app.

Processed files land directly in your device's Downloads folder. No temporary copies sit on a server, and no processing history is recorded anywhere.