Hebrew OCR — Extract Hebrew Text From Any Image
Upload an image containing Hebrew and get editable Hebrew text back in the original script. Suitable for printed documents, book pages, signage and screenshots where the text is locked in an image.
Drop your image here
or click to browse — JPG, PNG, WEBP, GIF…
Upload an image and click Extract Text
All readable text from the image will appear here.
What Makes Hebrew OCR Difficult
Hebrew letters are separate rather than cursive, which makes it easier than Arabic, but it brings its own difficulties. Several letter pairs are distinguished only by a small detail: ב and כ, ד and ר, ה and ח, or ו and ז, and at modest resolution these flip. Five letters take different final forms at the end of a word — מ becomes ם, נ becomes ן — so a misread form changes word boundaries. Vowel pointing (nikud) appears as dots and dashes beneath the consonants and is usually omitted in everyday text but present in religious, poetic and children's material, where it is small, dense and easily lost. Right-to-left flow also means lines mixing Hebrew with numbers or Latin terms can come back reordered.
Getting Accurate Hebrew Results
- Capture sharply enough to separate ב from כ and ד from ר — these pairs account for most Hebrew OCR errors.
- If your source has nikud, expect to proofread the points; the consonants extract far more reliably than the vowels.
- Check the five final letter forms (ך ם ן ף ץ) in the output, since a wrong form shifts where words appear to end.
- For Rashi script or handwritten Hebrew, treat the result as a rough draft rather than a transcription.
Who Uses Hebrew OCR
Religious and classical texts
Digitise printed Hebrew scripture, commentary and study material so passages become searchable.
Documents and correspondence
Extract text from Hebrew letters, forms and certificates without retyping right-to-left script.
Learners
Copy Hebrew off signs, menus and textbook pages to look up vocabulary you cannot type yet.
Hebrew OCR — Frequently Asked Questions
Is nikud (vowel pointing) extracted?▼
Consonants are extracted reliably; the vowel points below them are less consistent because they are small and closely spaced. For vocalised religious or poetic texts, plan to check the pointing against the original.
Does the output keep Hebrew script?▼
Yes. Text comes back in Hebrew characters with right-to-left ordering, not transliterated into Latin letters.
Will mixed Hebrew and English text work?▼
Yes, both scripts are read in one pass. Verify the sequence around numbers and Latin words, as mixing right-to-left and left-to-right runs on a single line can occasionally reorder segments.
Can it read Rashi script?▼
Partially. Rashi script is a distinct semi-cursive typeface used in traditional commentaries and its letterforms differ substantially from standard print, so accuracy drops considerably. Square Hebrew print gives much better results.
OCR in Other Languages
Or use the general image to text converter if your image contains several languages at once.