Merge PDF

About Merge PDF

A PDF looks like a stack of pages, but inside it is a set of numbered objects. Each page is a dictionary that points elsewhere: its content stream is object 12, its font is object 7, its image is object 9. The bookmarks, form definitions and document title hang off the catalog, outside any page. Merging two PDFs therefore isn’t a matter of concatenating bytes. The pages of the second file, and everything they reference, have to be carried into the first file and renumbered.

This tool uses the top file in the list as the base. That file is re-read whole and only its page order is rebuilt; pages from every other file are brought in with pdf-lib’s copyPages. That split decides what survives in the result.

The first file owns the document

Bookmarks (the Outlines tree), named destinations, page labels (the “i, ii, iii, 1, 2” numbering some viewers show), the initial view and the title in Document Properties all live in the catalog and the info dictionary, not on pages. copyPages copies a page and whatever that page references, so none of the catalog-level material from the second and later files comes along unless it is moved over separately. This tool moves the bookmarks over.

In practice the first file’s bookmarks stay at the top level of the merged file, and each later file’s bookmarks are nested under a new entry named after that file (it opens the file’s first page). Files without bookmarks get no entry. Bookmarks that used named destinations are pointed at the actual pages. The title, page labels and initial view still come from the first file, so put the main document — the contract, say, with exhibits appended — at the top.

Form fields and links

An interactive form field lives in two places. The widget you see is an annotation in the page’s Annots array, while the field’s name and value are registered in the document’s AcroForm dictionary. Copy only the page and the widget arrives without its registration; depending on the viewer, the box is drawn but can’t be filled in. This tool walks up from each copied widget to its top-level field and adds it to the first file’s AcroForm.

The awkward case is two fields with the same name. The PDF specification treats fields with the same fully qualified name as different views of one field that must share a value. Merge two copies of the same application form filled in by different people and you get two “FullName” fields; a viewer may show one person’s name in both, or mix them up when saving. So when a later file has a field whose name an earlier file already uses, this tool registers it under a numbered name — “FullName_2”, “FullName_3” — and lists the renames on the result screen. Fields that shared a name within one file were already one field, so they are renamed together. Calculations and scripts that refer to fields by name don’t know the new names; for such forms, flatten first so the answers become ordinary page content.

Internal links — “Go to Section 3” in a table of contents — point directly at a page object. Links inside the first file keep working because its page objects are reused as they are. When pages from the second and later files are copied, their internal links are re-aimed at the same pages as placed in the merged file, and named destinations are turned into the actual page. A link to a page you left out stops being clickable. Web links don’t depend on page objects and are unaffected.

Page size, rotation and fonts stay per file

Every page has its own MediaBox (paper size) and /Rotate value. Append a US Letter document (612×792 pt) to an A4 report (595×842 pt) and the result keeps both sizes; the Letter pages are about 6 mm wider and 18 mm shorter. That is deliberate. If a print shop wants one size throughout, run the merged file through the A4 fit mode of the Crop PDF tool.

Two files that use the same typeface don’t end up sharing one font. Most PDFs embed a subset — only the glyphs actually used — tagged with six capital letters and a plus sign, like “ABCDEF+Arial”. Each file’s subset contains different glyphs, so they stay separate objects and the merged file is roughly the sum of its inputs. It can come out a little smaller or larger because pdf-lib rewrites the file and packs small objects into compressed object streams.

Encrypted and scanned files

A PDF with an open password has its content encrypted, so its pages can’t be read and it can’t be merged; remove the password with the unlock tool first. A scanned page is a single large image and is moved across untouched, with no loss of quality. Scans add up quickly, though, so if a portal caps upload size, compress after merging.

Worked examples

How to use it

1. Select or drop the PDFs you want to combine.
2. Arrange the list. Put the file whose title and top-level bookmarks you want to keep at the top.
3. Click Create PDF, check the page count and download.

Frequently asked questions

Related guides

• How PDF Files Work: Objects, the Cross-Reference Table, and Why PDFs Are Hard to Edit