Which tool for which job
PDF chores fall into four groups: rearranging pages, shrinking the file, getting content in or out, and marking or locking the result.
- Combine several files into one attachment: Merge PDF.
- Take a document apart or pull out a chapter: Split PDF with ranges such as
12-15or5-. - Drop blank, duplicate or sensitive pages: Delete PDF pages.
- Pages open sideways or upside down: Rotate PDF.
- File too big for email or a portal: Compress PDF.
- Photos or scans into a PDF: JPG to PDF.
- Office documents into PDF: Word to PDF, Excel to PDF, PowerPoint to PDF.
- Content out of a PDF: PDF to Word for editing, PDF to text for plain text, PDF to JPG for images.
- A scan you cannot search: OCR PDF adds an invisible text layer.
- Finishing touches: Add page numbers, Watermark PDF, Sign PDF.
- Passwords: Protect PDF to lock, Unlock PDF to remove a password you know.
How a PDF is put together
A PDF page is a set of drawing instructions, measured in points of 1/72 inch: A4 is 595×842 points, US Letter 612×792. There are no paragraphs, only text placed at positions, which is why a PDF looks the same everywhere and is awkward to edit.
Two kinds of PDF behave very differently. A born-digital file, exported from a word processor or spreadsheet, contains real text: you can search, select and copy it, and a 50-page contract often weighs only 200 to 500 KB. A scanned file contains one picture per page. A colour page at 300 dpi is about 2500×3500 pixels and 1 to 3 MB, so a 20-page scan easily passes 30 MB. If Ctrl+F finds nothing, you have a scan.
The page tools here (merge, split, delete, rotate) copy pages as they are, without re-rendering, so text stays searchable and images keep their quality. Anything that belongs to the document as a whole, such as bookmarks and title or author properties, is not carried over.
PDF passwords come in two kinds. The open password encrypts the content; without it the file is unreadable. The permissions password only sets flags for printing or copying that readers are asked to respect. Protection here uses AES-256, the strongest method in the PDF standard.
Browser, server, or both
Most PDF tools run entirely in your browser and never upload anything: merge, split, delete pages, rotate, page numbers, watermark, sign, JPG to PDF, Word to PDF, PDF to text and PDF to JPG. Limits are 100 MB per PDF and 50 MB per image or .docx. Word to PDF lays the document out in the page and uses your browser's own "Save as PDF", so the text in the result stays real text.
Four tools offer both paths, and the choice changes the result:
- Compress PDF in the browser redraws every page as a JPEG at about 108 dpi. Excellent on scans, but text stops being searchable. The server path uses Ghostscript, which downsamples only the images (to about 72, 150 or 300 dpi) and keeps the text intact.
- PDF to Word in the browser rebuilds clean headings and paragraphs but drops images and tables. The server path uses LibreOffice, which keeps the look but places text in separate boxes.
- Protect PDF and Unlock PDF run qpdf in the browser for files up to 100 MB. For larger files the server path exists, and the password travels with the file for that one job.
Three tools are server-only, because no browser engine can do the work: Excel to PDF and PowerPoint to PDF use LibreOffice, and OCR PDF uses OCRmyPDF with Tesseract and saves the result as PDF/A. OCR jobs have a processing window of 15 minutes free and 60 minutes on Premium.
Server runs are free once per rolling 24 hours per device, for files up to 500 MB, without an account. Failed runs do not count. Premium removes the daily limit and accepts up to 4 GB. The uploaded file is deleted when the result is ready, and the result expires about an hour after the job was created, so download it right away.
Workflows that chain well
- An application pack: convert the letter with Word to PDF and the certificates with JPG to PDF, merge them in order, compress if needed, number the pages, and protect last. Encrypted files cannot be edited further, so the password always comes at the end.
- A stack of scanned paperwork: rotate sideways pages, delete blank backs, run OCR, then copy the text out with PDF to text.
- Numbering without the cover: split with
1, 2-60, number the second part starting at 2, and merge the cover back in front. - A year of bank statements: unlock them with the bank's password, merge into one file, and put your own password back on with Protect PDF.
To fit an upload limit by splitting, divide the file size by the page count: a 30 MB scan of 60 pages averages 0.5 MB per page.
When something goes wrong
- A file is refused as password-protected. Merge, split, compress, sign and OCR cannot read encrypted files. Remove the password with Unlock PDF first.
- Text extraction or PDF to Word comes back empty. The pages are images. Run OCR PDF first.
- The compressed file is no longer searchable. That is the browser path. Use the server option, or add a text layer back with OCR.
- The merged file is too large. Merging never shrinks anything. Compress afterwards.
- A certificate-signed contract shows as modified. Any tool that rewrites the file breaks a digital signature. Leave such files untouched.
- Word to PDF does nothing when you click. The browser blocked the print window. Allow pop-ups, then set margins to None and switch off headers and footers in the print dialog.