Transcribe printed English from a picture
Image to Text uses an English Tesseract OCR model. The image is processed in your browser. On explicit processing, its worker, engine and language data may download from third-party CDNs; first use is not guaranteed to work offline.
Prepare and recognize
- Choose an upright, clear image of printed English.
- Press Process. First use can take longer while runtime files and model data load.
- Select and copy the result, then compare it with the source.
- Choose Cancel recognition to stop and discard that operation's result.
Example: a photographed paragraph
A straight, well-lit crop of a paragraph is better input than a wide photo of a whole desk. Prepare it with Image Cropper and Image Rotator. Check common confusions such as 0/O, 1/l and lost punctuation; plausible text can still be wrong.
Can it read tables, handwriting or other languages?
Only the English model is loaded. Handwriting, complex tables, curved pages and decorative fonts can produce poor results. Output is plain text, not a reconstructed spreadsheet or layout-preserving document.
Can I select a scanned PDF directly?
Select an image instead. PDF to Images can render a suitable PDF page first. OCR here does not attach a searchable text layer back to the PDF.
The source must fit 25 MiB and 24 million pixels. Memory limits or blocked CDN/worker downloads can prevent recognition. Proofread exact numbers, names and measurements before using them. The Tesseract.js engine provides the underlying OCR; it is a transcription aid rather than a guarantee of accuracy.





