What Is OCR and How to Extract Text From Images and PDFs Free

Optical Character Recognition (OCR) is what turns a picture of text into text you can actually select, search, and copy — the difference between a scanned document being genuinely useful and just being a picture that happens to contain words.

What OCR actually does

A scanned page or a photo of a document is just pixels to a computer — there's no underlying text data, which means you can't search it, copy from it, or edit it. OCR analyzes the shapes in the image, recognizes them as characters and words, and outputs the result as actual text. It works best on clear, well-lit, reasonably high-resolution images with standard fonts, and struggles more with handwriting, low contrast, or heavily stylized text.

Extracting text from a photo or screenshot

Whether it's a photo of a whiteboard, a screenshot of an error message, or a scanned handout, the Image to Text (OCR) tool runs recognition directly in your browser and returns editable, copyable text — no need to retype what's already written down somewhere.

Extracting text from a scanned PDF

A PDF built from scanned pages — common for old contracts, government forms, and archived documents — has the same problem as a photo: it looks like text but isn't actually selectable. The PDF to Text (OCR) tool runs OCR across every page of a scanned or image-based PDF and extracts the text page by page, turning an unsearchable scan into something you can copy, search, or paste into another document.

Why on-device OCR matters

Because both tools run the recognition engine locally in your browser rather than uploading the image or PDF to a server, documents with sensitive content — IDs, medical records, signed contracts — never leave your device during the extraction process. This is a meaningfully different privacy model than the many free OCR tools that require an upload first, and it's worth checking for before running anything sensitive through an online OCR service. The first run does fetch the OCR engine itself over the network, but the actual image or document you're extracting text from is never part of that request.

Common scenarios

1. Digitizing a stack of old, scanned paper contracts into searchable text for an archive.
2. Pulling text out of a screenshot of an error message to paste into a support ticket or search engine.
3. Extracting a quote or passage from a photo of a book or printed page without retyping it.
4. Making an old scanned form's text searchable and copyable for the first time.

A note on accuracy

OCR accuracy depends heavily on image quality — a clear, high-resolution, well-lit scan will produce far more accurate results than a blurry photo taken at an angle. For best results, crop out unnecessary background and make sure the text fills a reasonable portion of the frame before running recognition. Straightening a tilted or rotated scan first also noticeably improves accuracy.

Try them

Extract text from a photo or screenshot with Image to Text (OCR), or pull text out of a scanned document with PDF to Text (OCR).