The proposal is drawn between the page thumbnails: a solid cut where the evidence is strong, a dashed one where it is a guess, and the reasons in each cut’s tooltip. Click the gap between two pages to add or remove a cut; blank pages are greyed out and left out unless you click to keep one. Each document is shown as a coloured band with its type and a name built from its own date, type and sender, such as 2025-01-31_Invoice_Acme_Supplies_Ltd, which you can type over. Download a ZIP with one PDF per document and an index CSV, or one PDF with a bookmark per document. Pages are copied, not re-rendered. Finding the cuts uses no job, and the file never leaves your device.
How it works
- Drop the PDF — Unstack reads it page by page on your device, with progress and a Cancel button
- Check the proposed cuts between the thumbnails, click to add or remove one, and fix any name or type
- Download one PDF per document with the index CSV, or one PDF with a bookmark per document
Questions people ask
How is this different from splitting every page or every N pages?
Split cuts by a rule you set, which works when every document has the same length. Unstack looks for where each document actually starts — page numbers, blank sheets, labelled numbers, letterheads, labelled dates, headers and footers — so a two-page invoice and a five-page statement in the same file come out as two documents. When you already know the page ranges, Extract pages or Split is quicker.
What if a cut is in the wrong place?
Click it to remove it, or click the gap where one belongs. Nothing is written until you download, the original PDF is never changed, and every cut shows its reasons, so a wrong guess is easy to spot. Dashed cuts rest on weaker evidence and are the ones to check first.
Does it work on PDFs that are not scans?
Yes. A page with text, such as an exported invoice or a statement downloaded from a bank, is read directly, which is faster and more reliable than OCR. A page with fewer than 40 characters of text is treated as a scan: its top and bottom strips are read with OCR on your device, which does not use the free OCR allowance.