How to Extract Text from an Image (OCR) for Free
Guide · Image Tools
OCR (optical character recognition) reads the pixels in an image and converts what it recognizes as text back into actual, editable, copy-pasteable characters — turning a photo of a page, receipt, or sign into real text instead of a picture that happens to contain words.
Extracting text in three steps
- Open Image to Text (OCR) and upload your photo or scanned page.
- The tool scans the image and returns the recognized text.
- Copy the result directly, or download it as a text file.
What makes OCR accuracy go up or down
- Resolution — text needs to be sharp enough to read; a blurry or very low-resolution photo produces more misread characters.
- Contrast — dark text on a light background is far easier to read accurately than low-contrast combinations or colored text on a busy background.
- Angle and skew — a photo taken at an angle rather than straight-on makes character recognition noticeably less reliable; straighten the image first if you can.
- Font — clean printed fonts read almost perfectly; handwriting, stylized fonts, and very small text are the hardest cases for any OCR engine.
OCR vs a PDF's built-in text
Not every PDF needs OCR — if a PDF was generated digitally (from Word, a website, or design software), its text is usually already selectable and extractable directly, no image recognition required. OCR specifically matters for scanned documents, photos of text, and PDFs made by photographing or scanning a physical page, where there's no underlying text layer to extract — only pixels.
After extraction, always proofread
Even accurate OCR occasionally misreads a character — a "0" for an "O", a smudge read as a stray letter. For anything that matters (a contract, an ID number, a quote you're going to publish), read the extracted text against the original image before using it, rather than trusting it blindly.