How browser PDF merge works (high level)

PDFs are boxes of pages

A PDF is not one giant image. It is a structured container: pages, content streams, fonts, images, and metadata. Merging usually means creating a new document and copying page objects from each input in order. That is why a good merge can keep vector text selectable instead of flattening everything into a photograph of the page.

In the browser, libraries such as pdf-lib parse those structures in JavaScript. When you drop two files into Omnipus Merge PDF, the library loads each PDFDocument, then appends pages into a fresh document. You can drag to reorder before the merge runs. The output is a brand-new file your browser downloads; the originals on disk are untouched.

What does not always survive

Bookmarks, some annotation types, and certain interactive form features may not copy perfectly. Omnipus keeps merges intentionally simple so the result opens widely in preview apps, phones, and printers. If you need advanced portfolio features preserved, a desktop PDF editor may still be the better tool.

Password-protected PDFs cannot be opened until unlocked. The browser cannot invent the password. Remove security in a trusted editor first, then merge locally. Encrypted attachments are a privacy feature — they are not something a website should bypass.

Memory and practical limits

Everything lives in the tab’s memory. Hundreds of light text pages are usually fine. Huge scanned books may stall a phone browser. When that happens, split the job into chapters, merge each chapter, then merge the chapter PDFs. That workflow still never requires uploading the book to a merge server.

If you care why Omnipus exists at all, the short version is this: merge is one of the most common reasons people visit sketchy upload sites. Doing it in a tab removes that excuse.

Worth knowing: a merge that copies pages is different from printing to PDF and combining images. Copied pages keep text selectable when the source had real text. If your inputs are scans, the merge still works — you just get a thicker file of images. Compress afterward if email size caps get in the way.

A quick privacy check after merging

Open the network panel once while you merge a throwaway PDF. You should not see the file body uploaded to Omnipus. You will see the page and library assets load — that is normal. If a third-party converter shows a large multipart upload the moment you click Merge, that is the old model. Prefer the local path for anything you would not paste into a random chat.