PDF Split

This tool does one narrow job: pick pages out of a PDF by number and download them as a new PDF. It suits the case where you want to send someone only a few pages of a long contract, report or scan. It produces a single file per run, so it will not cut a document into many pieces in one go; to split by chapter you run it once per range. It also cannot reorder pages, repeat a page or rotate anything.

Steps

1. Click the dashed box labelled "Click to select a PDF file" and choose a PDF from your device. The page shows "Reading file…", then "This file has N pages".
2. In the field "Enter the page ranges to extract", type something like 1-3,5,8-10.
3. Click "Extract & Download". The browser downloads a file named extracted-result.pdf (the Chinese interface names it 提取结果.pdf).

How the page range is parsed

The input is split on ASCII commas. Each piece is either one page number or two numbers joined by a hyphen. Take a 10-page file: 1-3,5,8-10 gives pages 1, 2, 3, 5, 8, 9 and 10. Typing 5,1,3 still yields 1, 3, 5, because the result is de-duplicated and sorted ascending. 10-8 is treated as 8-10, and 8-20 is clipped to 8-10 with the out-of-range part silently dropped. I ran the same parsing logic on a 10-page sample with these inputs and the results matched.

Supported input and known limits

• PDF files only, handled with the pdf-lib library. There is no artificial size cap, but the whole file is read into browser memory at once, so very large files can be slow or fail on phones and low-end machines.
• PDFs protected by an open password cannot be processed. The message shown is "Could not read this PDF file. Please make sure it isn't corrupted.", which does not mention encryption.
• Parsing is lenient: fragments it cannot understand are ignored without warning. The "No valid page numbers were found" error appears only when nothing at all is usable, so check that your input was not partly discarded.
• Only the ASCII comma and the plain hyphen work. With a full-width comma, "1，3" extracts just page 1, and "1-3-5" extracts only pages 1-3.
• The new file is built by copying the chosen pages into a fresh document. Page content comes along, but document-level structure such as the bookmark outline does not. The original file is never modified.

My result has fewer pages than I expected

Compare against the page count the tool shows, then look for full-width commas, en dashes or other unusual characters in your input. Those are dropped silently, so the output comes out shorter than intended.

Can it split a whole PDF into several files at once?

No. Each click exports one file. To split into chapters, enter each chapter's range in turn and export each time. Every download uses the same name, extracted-result.pdf, so rename them as you go.

It says it could not read my file

A corrupted file, something that is not really a PDF, and a password-protected PDF all trigger that same message. If you know the password, remove the protection in another program first.

Can I choose the order of the output pages?

No. Output follows the original page order, ascending, and a page listed twice appears once. Use the site's PDF Page Reorder tool if you need a different order.

Is my file uploaded?

No. The code makes no upload requests; reading, extracting and building the file all happen in your browser. The pdf-lib script module is loaded the first time you use the tool, and it is served with the site itself rather than fetched from a third-party server.