Extract text from image

1. Select or paste (Ctrl+V) an image.
2. Choose a language.
3. Click Extract text and copy the result.

How to get better recognition results

• Take the photo straight on, with the paper filling the image and no shadow from your hand or phone.
• Use even lighting: a spot near a window works better than flash, which can create a bright spot in the middle.
• Use sufficient resolution: small letters need more pixels. For scanners, 300 DPI is the most common setting for OCR.
• Choose the right language: the Portuguese model recognizes accents and cedilla; for mixed-language text, choose Portuguese + English.
• Crop out anything that isn’t text (photos, logos, table edges) with image cropper.

Text only or searchable PDF?

Text only gives you text to copy and a .txt file. Text + searchable PDF also creates a PDF with the original image and an invisible text layer on top. The document looks the same, but you can search for words (Ctrl+F) and select passages. It’s the best choice for archiving scanned documents. If you already have a multi-page scanned PDF, the PDF OCR tool is made for that.

What OCR doesn’t handle well

Tables may lose their column alignment (the text comes out line by line), and mathematical formulas and curved text often contain errors. Cursive handwriting is also less accurate. Check important numbers—account numbers, amounts, and dates—before using the extracted text.