The Smartest Free PDF Toolkit Online

Choosing the right PDF tool

Most PDF tasks fall into one of four groups, and knowing which group you are in makes the choice obvious. You are either changing the size of a file, changing its format, changing which pages it contains, or changing who can do what with it. Nearly every request people bring to a PDF toolkit is one of those four, and picking the wrong group is what turns a thirty-second job into an afternoon.

When the file is too big

Size problems almost always come from images. A page of plain text adds a few kilobytes, while a single scanned page can add several megabytes, which is why documents produced by a scanner are the largest files most people ever handle. A quick diagnostic is to divide the file size by the page count: anything above roughly 500 KB per page means images are responsible and compression will help substantially.

Compression works by reducing image resolution to what the page actually needs, so a scan commonly drops by 70 to 90 percent while a text document exported from a word processor may only shrink by 5 to 15 percent. If a file barely changes, that usually means it was already close to as small as it can usefully get, and the real fix is removing pages you do not need rather than compressing harder.

When you need a different format

Converting between PDF and an editable format means translating between two things designed for opposite purposes. A Word document describes flowing content that re-wraps to fit whatever it is opened in. A PDF pins every character to a fixed coordinate on a fixed page so it looks identical everywhere. Going from Word to PDF is straightforward because you are discarding flexibility you no longer need.

Going the other way is harder, because the structure was thrown away when the PDF was created. A PDF does not record where paragraphs begin or which lines belong to the same column, so a conversion has to infer all of that from position on the page. Documents that started life in a word processor convert back cleanly. Magazine layouts with multiple columns and floating text boxes are where the guesswork shows.

Scanned documents are a separate case entirely. A scan contains no text at all, only an image of text, which is why searching one finds nothing and converting one to Word produces a document with a picture in it. Optical character recognition solves this by reading the image and adding an invisible text layer behind the page. The document looks exactly the same afterwards and becomes searchable, selectable, and readable by screen readers.

When the pages are wrong

Page-level operations are the least risky things you can do to a PDF, because they rearrange content rather than re-encoding it. Merging, splitting, extracting, removing, reordering, and rotating all copy pages exactly as they were, so nothing is re-rendered and no quality is lost. Rotation in particular is worth understanding: the rotate button in most viewers changes only the display for that session, while a real rotation writes the orientation into the file so the correction follows the document everywhere.

Three of these overlap enough to cause confusion. Remove pages when you want the original document minus a few sheets. Extract pages when you want a new file containing only certain pages and do not care about the rest. Split when you want the document divided into several files and intend to keep more than one of them. Deciding what you want to be holding at the end makes the choice immediate.

When it needs protecting

A PDF can carry two different passwords and they are not equally strong. A user password is required to open the file at all and is enforced by encryption, so without it there is genuinely nothing to read. An owner password lets anyone open the document but asks their software not to allow printing or copying, which is a convention rather than a barrier. If content genuinely must not be read by the wrong person, it needs the first kind.

It is also worth separating three things people reach for interchangeably. Encryption controls who can open a document. Redaction removes specific content permanently, and covering text with a black box is not redaction, because the text is still in the file. Signing proves who approved a document and that it has not changed since. These stack rather than compete, and reaching for a password when the requirement was redaction leaves the sensitive part visible to everyone who legitimately opens the file.

The order that saves rework

When a job needs several steps, sequence matters more than people expect. Unlock any protected files first, since most operations cannot read an encrypted document. Fix page orientation next, because optical character recognition, cropping, and page numbering all depend on the page being the right way up. Run OCR before compressing, since recognition needs the full image detail that compression removes. Trim the document to the pages you actually want before merging, so every later step works on less material.

Leave compression and page numbering until the end. Numbers added before pages are merged or reordered end up in the wrong sequence, and compressing before merging means the savings get partly undone when shared resources are copied into the combined file. Working in that order means you only trade away quality once the free changes have run out.

How processing works here

Every tool runs server-side on dedicated engines rather than in browser JavaScript, which is why conversions complete in seconds rather than stalling on large files. The one deliberate exception is the document scanner, where edge detection and perspective correction run locally on your own device, so photographs of passports, contracts, and bank statements never leave your machine until you choose to create a PDF.

Files are limited to 50 MB each and are deleted automatically within one hour of upload. There are no accounts, no watermarks on any output, and no limit on how many times you use a tool. If a document matters, keep your own copy of the original before editing it here, since the version you upload will not be around to retrieve later.

Everything works in a normal browser on any operating system, including phones and tablets, with nothing to install and no plugin required. That matters most on a work machine where you cannot install software, and on a phone where the alternative is emailing a file to yourself so a desktop can deal with it. If a tool needs a setting, it is on the page rather than buried in preferences, and the defaults are chosen to be the right answer for the common case.