How to convert an image to text
- Pick the Language of the text first. The engine reads one language per run.
- Add your picture: drop it on Drop an image or PDF here, paste a screenshot with Ctrl+V (Cmd+V on a Mac), or tap Take photo on a phone.
- The first run downloads the engine and your language file (about 4 MB for English), and later runs reuse them.
- Check Mean word confidence, then fix slips in the editable Recognized text box.
- If the result is poor, tick Boost contrast or drag a rectangle on the preview to read only that part, then press Read again.
- Choose Keep line breaks or Join lines into paragraphs, then press Copy text or Download .txt.
How the text is recognized
OCR (optical character recognition) turns letter pixels back into characters. This page runs tesseract.js 7.0.0, the open-source Tesseract engine compiled to WebAssembly, so the work happens on your device. Before reading, the page enlarges small images, shrinks any image over 15 megapixels and adds a white margin, and Boost contrast turns the picture black and white. The result shows the mean of the engine's word scores.
Worked example with the sample image
Press Try a sample image. It is a 1,000 by 280 pixel screenshot of four lines of notes, read in English with default settings. It shows a mean word confidence of 95% (High), 22 words, 115 characters and this text:
Meeting notes, Tuesday 1. Send the signed form to Dana by Friday. 2. Book the room for 9:30 AM. Budget left: $2,450.75
Get better results from a photo or screenshot
- Screenshots are easiest because the letters are sharp. For a photo, hold the camera straight above a flat page in even light.
- Tesseract works best on scans of at least 300 dpi, per the Tesseract guide.
- Crop tightly, by dragging a rectangle on the preview or in the image cropper, so borders and logos are skipped.
Screenshot to text on Windows, Mac and iPhone
The Snipping Tool in Windows 11 and PowerToys Text Extractor (Win+Shift+T) copy text from a capture. Macs and iPhones have Live Text. Those are quicker for one line, so use this page for saved files, PDFs or another language. HEIC photos from an iPhone open here too, or convert them first with the HEIC to JPG converter.
Scanned PDF to text
Open the PDF, then use Pages to read to choose pages such as 1-3, 5. Each page is checked for its own text first: a page with real text is copied exactly and labeled, and a picture-only page is drawn larger and read with OCR. Tick Read every page as a picture to force OCR. A normal text PDF is faster in PDF to Text.
Languages and limits
Choose English, Spanish, French, German, Portuguese, Italian, Dutch, Polish, Turkish, Russian, Chinese (Simplified), Japanese, Korean, Arabic or Hindi. Handwriting, curved pages, low-light photos, tiny type and tables are weak spots. Join lines into paragraphs also glues a word split at a line break, so well-known becomes wellknown when a line ends at its hyphen.
Image to text questions people ask
How accurate is OCR on a photo or screenshot?
On a sharp screenshot or a flat, well lit page, accuracy is often above 95 percent of words. Blur, glare, small type and busy backgrounds lower it, so proofread anything below High.
Can it read handwriting?
Not reliably. The engine is trained on printed type, so cursive usually comes out as gibberish. Plan to retype handwritten notes.
Can it keep the layout of a table?
No. A table without borders comes out as one line per row, with a space between cells. Paste it into a spreadsheet and split the columns yourself. Bordered tables can come back nearly empty.
Can it convert a scanned PDF to text?
Yes. Open the PDF, choose the pages and press Read text. Pages that are only pictures are read with OCR, and up to 20 pages are read per run.
Sources: Tesseract documentation, improving output quality; tesseract.js project; Microsoft Support, Snipping Tool.