Which PDF Tool Should You Use? Merge, Split, Extract, Encrypt Explained

A PDF looks like one file, but it is really a container holding pages, fonts, images, vector graphics, forms, bookmarks, metadata, and security settings. What you want to do to it decides which tool to use, and the wrong choice costs you anything from a few extra steps to lost bookmarks, a broken signature, or a sensitive contract uploaded to a server you know nothing about. This guide is organized by what you are trying to achieve. It ties ToolGarden's PDF tools into one decision table and explains the right order for combining them and where each one stops.

First question: can this file leave your computer?

Before comparing features, judge how sensitive the file is. Contracts, ID scans, bank statements, medical records, and internal financials lose their protection the moment they land on someone else's server: how long it is kept, whether it is logged, whether it is scanned by third parties, you cannot verify any of it. ToolGarden's PDF tools (merge, split, extract, organize, encrypt, watermark, export images, convert to Word) run in the browser with JavaScript, and the file contents are not sent to a server.

You do not have to take that on trust. Open your browser developer tools, switch to the Network panel, choose a file and run the operation, then look for any request that carries the file. The download you get is a file generated in browser memory. The exceptions are the page's own static assets and, for PDF to Word, the one-time download of a local OCR model when a scanned page needs it. That model comes down to your browser; your file does not go up.

Decision table: from task to tool

One detail in that table is worth knowing up front. Merge, split, extract, and organize all work the same way internally: they create a brand-new PDF document and copy the selected pages into it. Page content (text, images, fonts, page rotation) comes along, but document-level structures such as the bookmark outline, form field definitions, document properties like title and author, and attachments do not automatically appear in the new file. Each section below returns to this.

Combining tools: order matters more than any single tool

Real work is rarely one step. A typical request is to turn several documents into one file for someone else and protect it. There is one hard rule for these workflows: encrypt last. An encrypted PDF needs its password to be opened, and the merge, split, and watermark tools cannot supply one, so they will report the file as possibly damaged or encrypted.

Workflow 1: a bid or application package

1. Use Organize PDF to remove blank or duplicate pages from the scans and fix any pages that are upside down.
2. Use Merge PDF to join the cover, main body, and credentials in the order the submission requires. Prefix file names with 01, 02, 03 so the list order is obvious.
3. Use PDF Watermark to mark every page with an English label such as CONFIDENTIAL.
4. Finally set an open password with PDF Encrypt and send the password through a different channel.

Workflow 2: slicing a long report for different departments

1. Use Split PDF and type 1-5; 6-18; 19-30 in the range box to get three files at once, downloaded as a ZIP.
2. If a department needs scattered pages, use Extract Pages and write 2,7,15-17.
3. When you need pages in a chat or document, export them with PDF to Image as PNG.

Workflow 3: scans to editable text

1. Run PDF to Word. It decides page by page: pages with a reliable text layer are extracted directly, pages without one go through local OCR, and anything still unsuitable falls back to a page image.
2. The result screen reports how many pages took each route, which tells you how far to trust the output.
3. Proofread numbers, names, and amounts in Word, since that is where OCR errs most.

Real limits of each tool

Merge, split, extract, organize

These four share the same input checks: the file must be a PDF, must not be empty, and must be 80 MB or smaller. An encrypted or damaged file produces the message that the PDF could not be read. Page numbers are always 1-based physical positions, meaning the Nth page in the file, not the number printed on the page. A book with roman-numeral front matter may show "iv (4 of 200)" in a reader; in the tool you type 4.

If you need bookmarks in the merged file, this approach is not enough: the new document does not generate an outline. Either finish that last step in a desktop PDF application, or replace the outline with a clear cover page and file naming. It is better to decide before merging whether readers will navigate by outline or read straight through.

Edit PDF

The editor supports rewriting existing text, adding new text, placing signature images, pen strokes, highlights, and cover boxes. Rewriting existing text is the feature most likely to surprise you. It edits the page content stream directly, so a rewrite is skipped when the text lives inside a form or graphics object, uses a multi-byte CID font, or when the original font has no glyph for the characters you typed (for example, replacing text in a Latin font with Chinese). The tool tells you how many rewrites could not be applied and why, and everything else is still written to the file. Always zoom in and check rewrites on important documents.

Watermark

The watermark tool offers center, tile, and diagonal layouts with adjustable size, opacity, rotation, and color; opacity is clamped to a minimum of 0.05. It embeds the Helvetica standard font, so write the watermark in Latin letters and digits. CJK characters cannot be encoded by that font. If you need a non-Latin watermark, create an image elsewhere and place it on the pages with Edit PDF.

Encrypt and decrypt

Encryption supports an open (user) password and an optional owner password, with checkboxes for allowing printing, copying text, modifying, and annotating. If you leave the owner password empty it reuses the open password. The key point is that permission flags are declarations written into the file: well-behaved readers honor them, but they are not strong protection. What actually stops a stranger is the open password. Our dedicated encryption article goes into the details.

Image export and PDF to Word

PDF to Image renders each page on a canvas with pdf.js. Choose PNG or JPG (JPG quality is 0.92) and a scale from 1 to 4 in steps of 0.5; a higher scale is sharper and heavier, and multi-page results can be bundled into a ZIP. PDF to Word outputs a DOCX that keeps font sizes, basic styling, and images, and scanned pages can be recognized in English, Simplified Chinese, Traditional Chinese, or Japanese. Both run locally.

A full example: sending five items to a client

Suppose you have a 28-page contract, a 6-page quote as attachment A, 12 pages of credentials as attachment B, and a 3-page scanned signature section whose page 2 is blank. The goal is a 48-page file for the client to review, protected against casual forwarding.

Step                          Pages            Notes
1 Organize (signature scan)   3 -> 2           delete blank page, outputs xxx-organized.pdf
2 Merge (four files, ordered) 28+6+12+2 = 48   merged page count = sum of inputs
3 Watermark                   48 -> 48         English text CONFIDENTIAL, diagonal, opacity 0.2
4 Encrypt                     48 -> 48         open password, restrict modify and annotate

The page count after step 2 is the cheapest check there is: the merge result shows the total, and if it is not 48, a file was missed or added twice. After step 3, flip to the first, a middle, and the last page to confirm the watermark is everywhere. After step 4, reopen the file in a different reader to confirm it asks for a password and that printing and copying behave as intended.

Common pitfalls and how to diagnose them

When not to use these tools

• Formal archival files that must keep bookmarks, form fields, PDF/A conformance, or accessibility tags (a structure tree) need desktop software that preserves them.
• True redaction, where blacked-out content must be unrecoverable, is a different job: deleting a page or covering text with a box does not remove the text underneath. Export, then check whether the text can still be selected and copied.
• Single files over 80 MB, or documents with thousands of pages, can exhaust browser memory; process them in batches.
• Legally meaningful e-signatures or timestamps call for a dedicated signing service, not a signature image placed on a page.

Summary

Choosing a PDF tool comes down to three sentences. Decide whether the file may leave your computer, and if it may, compare features. Use merge, split, extract, and organize to change which pages exist, edit and watermark to change what is on a page, and image export or Word conversion to change the file's form. Encrypt last, and sign even later. Keep that order and most everyday PDF jobs take a few minutes.