Documents—
Pages out—
Bookmarks—
Size—
Combining PDFs is treated as a solved problem, and the page count always comes out right — which is why nobody checks the rest. Merge a set of chaptered reports with almost any tool and the bookmark pane of the result is empty, the filled-in form fields have stopped being fillable, and every link that pointed at another page now points nowhere. This merger rebuilds the outline instead of discarding it, offers to flatten form values so they survive, and prints what the merge costs before you download anything. It also never sees your documents: the whole thing runs in this tab.
Step by step
01Add the documents
Drop them anywhere on the page or choose them from the inspector. Any number, any size — there is no upload, so there is nothing to queue and nothing to cap.
02Put them in order
Each document is a row with its page count, paper size, and what it carries. Move rows with the arrows, or sort every file by name at once from the topbar.
03Decide what happens to the bookmarks
Keep all rebuilds every document's outline under an entry named for the file. One per file writes a single entry each. None writes no outline — which is what merging without this control does silently.
04Read the report, then download
Bookmarks written, form fields flattened or left inert, links kept against links broken, and whether the pages are all the same size. Then take the file; it is already built.
A worked merge
The two sample documents this tool opens with. One is a seven-page chaptered report on A4 with a landscape page in the middle; the other is a two-page response form on US Letter with two fillable fields. Both numbers below are what the tool reports on them.
Given
The report this merge prints
Pages 9
Bookmarks 11 written (9 found in the sources, plus one entry per file)
Form fields 2 flattened, 0 left inert
Links 0 kept, 0 broken
Page size A4 · A4 landscape · Letter ← three sizes, worth knowing
Output 7.4 KBNine pages is the number every merger agrees on. The eleven bookmark entries are the difference between a document you can navigate and a 9-page scroll, and switching the bookmark control to None reproduces exactly what a merger that never mentions bookmarks hands you. The two flattened fields are the second difference: copied as-is they would still be drawn on the page and would no longer accept typing, which reads as a broken document rather than a converted one. And three page sizes in one file is not a defect — but it is the reason the printout comes out with two different margins, and it is better to learn that here than at the printer.
What a merge does to a document
Bookmarks are a separate tree, and copying pages does not copy it
A PDF outline lives in the document catalog, not on the pages, and it points at pages by object reference. Copy the pages into a new document and every one of those references is stale, so the standard behaviour is to write no outline at all. Rebuilding it means reading each source's tree, renumbering every entry to its new page, and writing the whole thing back.
A copied form field looks fillable and is not
The boxes you see on a form page are widget annotations, and those are copied with the page. What is not copied is the AcroForm dictionary that ties them to the document. The merged file therefore shows every field and accepts nothing — which is why flattening the values into the page first is usually the right answer, and why it is the default here.
Links to a URL survive; links to a page do not
An external link stores a string, so it comes through a merge untouched. An internal link stores a reference to the page it jumps to, and copying that reference pulls a stale copy of the old page along rather than pointing at the new one. Any merger that does not rewrite them leaves the link there and broken.
Mixed page sizes are the most common ugly merge
A4 is 210 by 297 millimetres and US Letter is 216 by 279 — wider and shorter. Merged into one file they print with different margins and scroll at different widths. Nothing is lost, but the result looks wrong, and the cause is invisible unless something counts the distinct sizes.
Attachments are not on the pages either
Files embedded in a PDF hang off the document's name tree, in the same place as the outline. A page-level copy leaves them behind, so a merged file quietly loses the spreadsheet that was attached to the original. This tool counts them and says so.
Order is decided by you, not by upload timing
Services that merge on a server frequently order by the sequence uploads completed, which on a slow connection is not the order you chose. Here the list is the order, the rows move, and sorting by file name is one click — useful when the files are named part-1, part-2, part-10.
The judgement call
What the sources are decides this, not what the output is for.
Chaptered documents that each have their own outline
Keep all
Every entry is renumbered and nested under its file, so a 300-page merge stays navigable at the same depth the originals were.
A pile of single-topic files — invoices, scans, certificates
One per file
The sources have nothing to preserve, and one entry per file turns the merged document into a table of contents of its parts.
Files whose outlines are stale or wrong
One per file
Rebuilding a bad outline preserves the badness. A single entry per file is a clean structure you can leave as it is.
A merged file going straight into another tool
None
Some downstream pipelines rewrite the outline anyway. Writing none is smaller and cannot conflict.
You are not sure
Keep all
It is the only mode that cannot lose anything: entries the sources carried come through, and files without an outline still get an entry of their own.
Reference
FAQ
Use a merger that runs in the browser rather than on a server. The Tinapps merge PDF tool reads each document with pdf-lib inside the page, builds the combined file in memory, and hands it back as a download — no request carries the file, so it works with the network disconnected and there is nothing on a server to delete afterwards.
Because a PDF outline is stored in the document catalog and points at pages by object reference, not by page number. Copying pages into a new document invalidates every one of those references, so most mergers write no outline at all rather than fix them. The Tinapps merge PDF tool reads each source's tree, renumbers the entries to their new pages, and writes it back.
They stop being fillable unless something is done about it. The field boxes are widget annotations and travel with the page, but the AcroForm dictionary that makes them interactive does not, so the merged file shows every field and accepts no input. Flattening the fields first — painting their current values permanently into the page — is what keeps the content, and it is the default in the Tinapps merge PDF tool.
Not in the Tinapps merge PDF tool. iLovePDF limits free merges to 25 files and 100 MB and Smallpdf meters free tasks per day, because both are paying for server time. This one uses your own processor and memory, so the practical ceiling is what the browser tab can allocate — hundreds of megabytes on a desktop, less on a phone.
No. Merging copies page content streams, fonts, and images across byte for byte; nothing is re-rendered or re-compressed, so text stays sharp and images are identical to the originals. A merged file is usually slightly smaller than the sum of its parts because shared resources are stored once.
Yes — a PDF stores a size per page, so a single file can mix A4, Letter, and landscape pages without any problem. It will print and scroll unevenly, which is why the Tinapps merge PDF tool lists the distinct sizes in its report rather than letting you find out at the printer.
No, and neither can the upload-based services without the password. The page content of an encrypted PDF cannot be read until it is decrypted, so the only path is to open the file in a PDF viewer, save an unprotected copy, and merge that. The Tinapps merge PDF tool says exactly this rather than failing with a generic error.
The merged file gets one set of document information rather than several, since a PDF has only one. The Tinapps merge PDF tool sets the title from the first document or from whatever you type, and identifies itself as the producer; the individual sources' authors and keywords are not carried over, because there is nowhere to put more than one.
The documents are parsed and rewritten entirely in your browser with pdf-lib. Nothing is uploaded, nothing is stored, and the merge keeps working with the network disconnected.
Keep going
PDF tools
The rest of the category — split, rotate, extract, and convert, all in the browser.
Paper sizes
What A4, Letter, and Legal actually measure, and which is closest to which.
Image compressor
Shrink the pictures before they go into a document rather than after.
MIME types
Why a PDF is application/pdf, and what a browser does with the header.