Compare two PDF files

How the comparison works

The tool reads the text of each page with pdf.js, splits it into words, then looks for the longest run of words shared by A and B. Words outside that run are differences: red strikethrough in A for what disappeared, green in B for what was added.

\[ \text{similarity} = \dfrac{2 \times \text{shared words}}{2 \times \text{shared words} + \text{removed} + \text{added}} \]

Pages are compared by number: page 3 of A against page 3 of B. If you inserted a page in the middle, the following pages will look shifted; line the pages up first with the Organize tool.

Reading the result

What the comparison cannot see

Only the text is compared. A changed image, color, font or position does not show up. For scans, which hold no text, run OCR first. Changed numbers (a price, a date) stand out clearly, which makes the tool handy for proofreading a contract or a quote.

Compare PDF files: real-world uses

When you need to compare PDF files, the hard part is spotting one changed word in ten identical pages. The tool does that for you. Here are the cases where it helps most.

• Contracts: first, you confirm the signed copy holds exactly the clauses you negotiated.
• Quotes: a changed price or quantity jumps out in red and green right away.
• Homework: also, you see what a classmate or a tutor changed in your draft.
• Forms: when an agency updates a form, you spot the new wording fast.
• Reports and essays: after a review, you check that only the announced edits happened.
• Exports: for example, you compare yesterday's file with today's to prove nothing moved.

If you compare large PDFs, close your other tabs first. Reading the text takes memory, so a lighter browser works faster.

In addition, the result does not depend on file size: a two-page PDF and a fifty-page PDF follow the same method. However, a longer document takes longer. So wait until the calculation ends before you read the percentage.

How to compare two PDF files step by step

1. First, drop the old version in zone A.
2. Next, drop the new version in zone B.
3. Then click Compare and wait a few seconds.
4. Finally, read the similarity percentage and scroll to the pages with marked words.

If a file asks for a password, the tool prompts you before it starts. So type it in: it stays in your browser. Then the comparison starts. You do not need to click anything else. Meanwhile, the tool keeps both files on your device. Afterward, you can pick another pair, so a new check takes seconds. When the report looks right, download it and attach it to an email or a case file.

Common mistakes when you compare PDF documents

Then, if you need to fix a detail, open the file in the edit PDF tool. To reorder pages, use organize PDF. For a scan, run OCR before you compare.

For example, a ten-page contract with one page added at position 4 looks fully changed. Yet only one page differs. So before you panic, look at the structure. Then move the pages back. Finally, run the comparison again. The result becomes readable.

Tips for a reliable comparison

• Always compare files of the same kind: two exports from one program, or two drafts of one contract.
• That way, you avoid noise from a very different layout.
• Keep the old version in A. As a result, red shows what disappeared and green shows what arrived.
• Read the percentage with care, because 98% on a long contract can hide one key clause.
• Before you share a sensitive file, use PDF redaction to hide private data.

Finally, when a passage looks odd, check it against the source document. A struck-through word always deserves a second look.

Last, keep a copy of the report with both file names. That way, you can tell which version replaced the other. In addition, this simple log ends doubts when several people review the same document. If a disagreement comes up, you point to the exact difference, word by word.

• So, put a date in your file names.
• Next, always compare the newest version with the one before.
• Then write the percentage in your log.
• Finally, reread every red word before you sign off.

However, a high percentage does not prove the files match. It only measures text. So keep an eye on images, stamps and signatures. Likewise, glance at the footer, because a changed page number is text too. Besides, a short manual scroll costs nothing. Next time, name each file clearly. Then you never mix up A and B. Also, work from copies, so the originals stay untouched. After that, delete the copies you no longer need.