Image to text (OCR)

Optical Character Recognition (OCR) turns text inside an image into editable, searchable text. Instead of retyping a screenshot, a photo of a whiteboard, a scanned letter or a slide, you can extract the words in seconds. This tool performs OCR in your browser, so documents with personal or business information are not uploaded.

How to use it

1. Select or drop an image, such as a PNG screenshot or JPEG photo.
2. Choose the language of the text if options are available.
3. Wait for recognition to finish.
4. Review the extracted text, correct any errors and copy it.

Getting the best results

• Resolution: text should be reasonably large; around 20 to 30 pixels tall per line of text works well. Tiny text in a low-resolution image is hard to read.
• Contrast: dark text on a light background gives the best accuracy.
• Straight and flat: photograph documents straight on, avoiding curves, shadows and glare.
• Crop to the area with text so other elements do not confuse recognition.
• Printed text is recognised far better than handwriting.

Typical uses

• Copying error messages from screenshots into a search engine.
• Digitising receipts and invoices for expense reports.
• Extracting quotes from slides or book pages.
• Making scanned documents searchable.

Limitations

OCR outputs plain text. Layout such as columns and tables may be flattened, and characters like 0 and O or 1 and l can be confused. Always proofread numbers, names and codes before using them.