PDF Merge & Split

What this tool does

It does two jobs. Merge joins several PDFs, in the order shown in the list, into one new PDF.Split takes one PDF and either collects the pages you name into a single new file or turns every page into its own file. Typical uses are bundling a cover letter, a form and a few supporting documents into one attachment, or sending a colleague only chapters 3 and 4 of a long report. It does not rotate, reorder pages inside a file, edit content, compress, or convert other file types into PDF.

Where your files are processed

The page uses the open-source pdf-lib library (version 1.17.1 in this project). The library is built into the page's own JavaScript when the site is published; it is not loaded from a third-party CDN at the moment you use the tool. When you choose a file, the code reads it withfile.arrayBuffer(), opens it with PDFDocument.load, copies pages into a new document with copyPages, writes the result with save, and hands you the bytes as a downloadable Blob. Nothing in this page's code sends the contents of your PDFs over the network, so nothing is uploaded or stored. If you want to see this for yourself, open your browser's developer tools, switch to the Network tab and run a merge: you will not see a request that carries your file. The site-wide parts of the page (such as ad and analytics scripts) run separately from the PDF code and are not given your documents. Closing or refreshing the tab discards everything held in memory.

How merging behaves

Files you pick are appended to the list, even if you choose them in several rounds. Each row shows the file name and size and has up, down and remove buttons; the merge button turns on once there are at least two files. The output is always called merged.pdf, and the status line reports how many files and pages went in and how big the result is. Adding the same file twice puts it in twice, which is handy for repeating a signature page or a blank separator. Because pages are copied rather than re-rendered, text stays selectable and images are not recompressed: quality is unchanged, and so is size. Expect the merged file to be about the size of the inputs added together. If the result is too large to email, shrink the source files first, for example by recompressing embedded photos with the image compressor before the images are placed in the document.

Page ranges for extracting

"Extract Selected Pages" builds one extracted.pdf from the numbers you type. Separate items with commas and write a run of pages with a hyphen, as in 1-3,5. Numbering starts at 1 and follows the order you type, so 3,2,1 gives you those three pages in reverse. A reversed range such as 5-3 is treated as 3 to 5. Page numbers beyond the end of the document and any piece the parser cannot read are skipped, and the line under the input box tells you how many pages will actually be extracted and what was ignored. The table below was calculated while the site was built, by running the same rules against a 10-page document.

Two rows deserve a closer look. 2,2,4 repeats page 2, and the output really does contain it twice. 8-12 is cut off at page 10 rather than rejected. The open-ended 7- is not a valid range at all, so it is ignored instead of meaning "7 to the end"; write 7-10. "Split Every Page Individually" makes N separate one-page PDFs from an N-page file and lists a download button for each (page-1.pdf, page-2.pdf and so on). Nothing is saved automatically: you click the buttons you want. Large documents take a while because every page is built as its own document, and a browser may ask whether to allow several downloads from one page.

Worked examples

• Merging an application packet. A 12-page contract, a 3-page ID copy and a 1-page bank letter merged in that order give a 16-page file (12 + 3 + 1). The status line reads "3 files, 16 pages total", followed by the output size.
• Extracting from a 30-page report. Typing 1-3,10,25-30 yields 3 + 1 + 6 = 10 pages. Typing 25-35 on the same file keeps only pages 25 to 30, which is 6 pages, and the preview tells you that part of the range was out of bounds.
• Reordering with the range box. To move the last page to the front of a 5-page file, type5,1-4. The result is page 5 followed by pages 1 to 4, again 5 pages.

Limits worth knowing before you rely on it

• Password-protected or encrypted PDFs. There is no password prompt. pdf-lib refuses to open an encrypted file with its default settings, so the tool shows a failure message. In a merge, one failing file stops the whole run and no output is produced. If you legitimately know the password, save an unprotected copy from a PDF viewer first.
• Size and count. The code sets no cap on the number or size of files. But every selected file is read fully into memory, and merging builds the output in memory too, so the real ceiling is your device. Phones and older laptops can slow down or run out of memory with a few hundred megabytes; merge in smaller groups if that happens.
• What may not survive. Pages are copied into a brand-new document. Bookmarks, document properties such as title and author, and some interactive items (form fields, links, annotations) are not guaranteed to carry over. Open the result and check anything important.
• Damaged or unusual files. A PDF with a broken structure, or one produced by unusual software, may fail to load. Re-saving it with "Print to PDF" in another program often cures this.
• Only PDFs. The file picker accepts PDF files. Word documents and images have to be turned into PDF elsewhere first.
• Per-page downloads are temporary. The buttons created by "Split Every Page Individually" exist only on the current page view and disappear on reload or when you pick another file, so download what you need right away.

After you have your file

Open the result in the PDF reader to confirm the page order and that nothing is missing before you send it. If you need the text of a document counted or checked, theword counter works on pasted text. When a merged file has to go to a site with an upload limit, check its size first and, if necessary, split it into two parts with the range box instead of squeezing it. And if the file contains personal data, share it only through a channel you trust, since this tool stops protecting it the moment you download it.

Frequently Asked Questions