PDF is one of the few everyday formats where the usual answer is "upload it somewhere" — and the files in question are usually contracts, tax paperwork or medical reports. These five take the other route: the document is parsed and rebuilt inside your browser and saved straight back out.
Which tool should I use?
- Merge PDF
- For joining two or more PDFs into one, with arrow buttons to reorder them first. Know in advance that pages are copied into a brand-new document, so bookmarks, form fields and document properties from the sources do not come along. The text itself is fully preserved and stays searchable.
- Split PDF
- For pulling the pages you need out of a document; ranges accept the
1-3,5,8-form, or click thumbnails to select. Two things to expect: the output is always a single file rather than one per page, and selected pages always come out in their original order — entering3-4,1yields pages 1, 3 and 4. - Rotate PDF
- For scans that arrived sideways, or the two landscape pages in the middle of a report; each click on a thumbnail turns it 90 degrees clockwise. The angle is added to whatever rotation the page already declared, so a page marked 90 degrees becomes 180 after one click. Unlike merge and split, this one edits the original document and saves it back, which is why bookmarks and form fields survive.
- PDF to Image
- For dropping page content into a slide deck, or feeding a system that only accepts images. Pages are rasterised to PNG or JPEG at a scale of 1 to 4, roughly 72 to 288 dpi. The output is a bitmap with no text layer, so the words in it can no longer be searched — use split when the text has to stay live.
- Image to PDF
- For collecting photographed receipts, whiteboards or signature pages into one sendable file, with several images selectable at once and reorderable. Two layouts are offered: match the image, or A4. Matching treats pixels as PDF points, so a 4032 × 3024 phone photo produces a page roughly 56 by 42 inches — choose A4 if it is going to a printer.
Shared limitations
- The limits are 50 MB per file, 100 MB combined for a merge, and 500 pages; long documents consume noticeable memory just generating thumbnails. Password-protected PDFs are refused outright — nothing here tries to work around the protection, and there is no password field.
- The pdf.js worker that parses your document is served from this site's own
/pdf.worker.min.mjsrather than an external CDN, and the file never leaves the browser at any point; the page separately loads Google's advertising code, an unrelated path that never sees your document.
FAQ
Why does my PDF fail to open with an encryption error?
A PDF can hold two kinds of password: one required to open it, and an owner password restricting printing or editing that never prompts the reader — which is why people are often unaware their file has one. Both are refused here rather than bypassed. Re-saving or printing to PDF from a reader that already opens the file normally produces an unprotected copy.
Does Chinese text survive when I rasterise a PDF?
Yes. Rasterisation goes through pdf.js, which needs an extra character-mapping table for the predefined CJK encodings, and this site ships that table — so a document using those encodings without embedded fonts still renders. The same set of assets covers the CCITT, JBIG2 and JPEG 2000 decoders that black-and-white scans depend on. Split and rotate redraw nothing, so neither was ever affected.