Is my PDF scanned or digital? The 10-second test

How do I tell if a PDF is scanned or digital?

Open the PDF and try to highlight a single word with your mouse. If the word highlights like text in a web page, there is a text layer. If you only get a rectangle over part of the page, or nothing happens, the page is a picture.

A second test: press Ctrl+F (Cmd+F on a Mac) and search for a word you can see on the page. If the viewer finds it, there is a text layer. If it says no matches for a word that is plainly there, the page is a picture.

Run the test on a few pages, not only the first. It takes about ten seconds and needs no software beyond the viewer you already use.

What are the three kinds of PDF?

There are three, and the difference decides what you can do with the file.

• Digital PDF. Made by software, such as a bank's website, Word or a design tool. Every letter is stored as text, so you can select it, search it and copy it.
• Scanned PDF. Made by a scanner or phone camera. Each page is one picture. It looks like a document, but the computer sees no letters.
• Scanned PDF with a text layer. A scan that someone already ran through OCR (optical character recognition, software that reads text from pictures). The picture stays on top and the recognised text sits hidden underneath. You can search it, but the hidden text may contain mistakes.

What does each kind need?

A digital PDF needs no OCR to get its text. Copy it out, or use a PDF-to-text converter. A scan needs OCR, which reads the picture and writes the letters it finds. If you want the PDF to stay looking the same but become searchable, use a searchable PDF, which adds an invisible text layer under the page image.

The PDF to text tool gives you plain text. The make a PDF searchable tool gives you a PDF with the hidden layer. If you are unsure which you need, the searchable PDF vs plain text answer explains the choice.

Why does a digital PDF still paste garbage?

Because a text layer is only as good as the font behind it. Some PDFs, especially older ones, use fonts that store letters under the wrong codes. The page looks perfect, but the hidden text is nonsense. Copying gives you gibberish, or boxes, or missing vowel signs.

This happens a lot with Indian-language PDFs made with pre-Unicode fonts. Our guide to garbled Hindi from PDFs shows an example. The fix is to read what you can see, not what is hidden, which is what OCR does.

What if my PDF mixes both?

Some files are a mix, such as a typed contract with a signed scan stapled on, or a report with a pasted screenshot of a table. The text layer can be present on page 1 and missing on page 6. Test a word on every kind of page you care about, and test inside tables and images, not only in the body text.

FastOCR's PDF OCR reads the page image for every PDF and does not trust the existing text layer. That makes mixed files simple, since each page is read the same way, and it avoids broken text layers and old fonts.

When is this test not enough?

The test tells you whether text exists, not whether it is right. A scan with a bad hidden layer passes the highlight test and still copies wrongly. Paste a few lines into a plain text file and compare them with the page. For documents where every figure matters, read the results against the original.