The short answer
Drop the image onto the tool below, or copy a screenshot and press Ctrl+V (Cmd+V on a Mac). Check that Language of the text matches the image, wait a moment while it reads, then press Copy text or Download TXT. Several images can go in at once, and each appears as its own result.
Recognition runs in your browser with Tesseract, a long established open source OCR engine. The language data you select is downloaded once (1 to 3 MB) and then works offline.
Try it here: Image to Text (OCR)
Open the full tool pageHow OCR works and why the language matters
OCR software first cleans the image, finds lines and words, and then matches each word against what it knows about the shapes of letters and common words in a language. That last step is why choosing the right language makes such a difference. With English selected, a Turkish ş or a German ß has no match and turns into the closest English letter; with the right language the accents and special characters come out correctly.
The tool supports 20 languages, among them English, Spanish, Portuguese, French, German, Italian, Dutch, Polish, Turkish, Russian, Ukrainian, Arabic, Hindi, Chinese (simplified and traditional), Japanese, Korean, Indonesian, Vietnamese and Thai. It reads one language per pass, so for a page that mixes two languages, run it with the main one.
Copying text on your phone or computer
Most devices can now copy text from an image without any extra app. These built in options are handy for a quick line; a dedicated OCR tool is better for whole pages, many images, or languages your device does not cover.
- iPhone and iPad: open the photo or screenshot in Photos, touch and hold a word, adjust the selection and tap Copy. The Live Text button in the corner selects all text at once. It works in many supported languages on recent iOS versions.
- Android: open the image in Google Photos and tap Lens, then choose the text and Copy. Some phones also offer text selection right from the screenshot preview.
- Mac: Live Text works in Photos, Preview and Safari on recent macOS versions. Hover over the text in an image and select it like normal text.
- Windows 11: in the Snipping Tool, take or open a screenshot and use Text actions to copy all text. Microsoft PowerToys adds a Text Extractor shortcut for any part of the screen.
- Any device with a browser: open this page, choose or paste the image, and copy the result. This also works for scanned PDFs exported as images and for languages the built in features do not handle.
Getting accurate results
OCR accuracy depends far more on the image than on the software. A sharp screenshot is often read with almost no errors, while a dim photo taken at an angle can produce nonsense. These are the factors that matter most.
| Factor | What works | What causes errors |
|---|---|---|
| Light | Even daylight or a bright room, no shadows | Dim light, a phone shadow across the page, flash glare |
| Angle | Phone held flat and straight above the page | Shooting at a slant, curved book pages |
| Resolution | Letters at least 20 pixels tall | Tiny text in a wide photo of a whole desk |
| Contrast | Dark text on a light background | Gray on gray, text over photos or patterns |
| Focus | Tap to focus, hold still | Motion blur, out of focus text |
| Cropping | Only the text area | Pictures, logos and borders around the text |
Handwriting, tables and other limits
- Handwriting: the engine is built for printed text. Very neat block capitals may work; cursive and notes usually come out wrong. For handwriting, phone features such as Live Text or Google Lens tend to do better, but always check the result.
- Tables and columns: line breaks and paragraphs are kept, but table cells come out as text lines, not as a spreadsheet. For receipts and invoices, copy the text and tidy the columns afterwards.
- Stylized fonts: decorative, very thin or outlined letters on posters and logos are often misread.
- Rotated photos: the orientation saved by the camera is applied before reading, but a page photographed upside down without that information should be rotated first.
- Low confidence: when parts were hard to read, the tool says so. Retake the photo closer and in better light rather than editing many wrong words by hand.
Tips for scans, screenshots and long documents
Screenshots give the best results because the letters are perfectly sharp. If you can choose, take a screenshot of the text on screen instead of photographing the monitor.
For paper documents, a scanner or a scanning app at 300 dpi gives cleaner results than a quick snapshot. Small images are enlarged automatically before reading, but they cannot gain detail that was never captured.
For a multi page document, photograph each page separately and drop them all at once. Each page becomes its own result, so you can copy them in order into one document.