Multilingual OCR (Arabic, Russian, Chinese…)

Recognise text mixing up to three languages, including Latin, Cyrillic, Greek, Arabic, Hebrew, Hindi, Chinese, Japanese and Korean.

AI Private · runs on your device Free
Runs locally in your browser AI processing happens on your device. Your files are not uploaded for inference.
AI model
…
Download size
… ·
Device
Checking browser support…

    Different scripts, one tool

    A menu in English and Arabic, a product label in English and Russian or a document in Chinese and English: Multilingual OCR lets you choose the languages in the image and recognise up to three at once.

    Supported scripts

    • Latin: English, Turkish, German, French, Spanish and more
    • Cyrillic: Russian, Ukrainian, Bulgarian
    • Greek
    • Arabic and Persian (right to left)
    • Hebrew (right to left)
    • Hindi (Devanagari)
    • Chinese, Japanese and Korean

    Right-to-left text

    Languages such as Arabic, Persian and Hebrew are shown right to left in the result. Arabic support is included here rather than as a separate tool because it uses the same engine and workflow.

    Only the languages you need

    For each selected language Tesseract downloads only that language model and caches it in your browser. Your image isn't uploaded.

    How to use Multilingual OCR (Arabic, Russian, Chinese…)

    1. Upload an image containing text.
    2. Choose the languages in the image (up to three).
    3. Wait for the language models to download and recognition to finish.
    4. Copy or download the recognised text.

    Why use this tool?

    Digitise text in different scripts in one tool, for free and privately.

    FAQ

    Why should I only pick the languages in the image?

    Every extra language increases the download size, slows recognition and raises the chance of errors. Only select languages that really appear.

    Does it recognise Arabic handwriting?

    No, it's designed for printed text; results for handwriting and decorative calligraphy are poor.

    Does it read vertical Japanese text?

    Tesseract works best with horizontal text; accuracy can drop for vertical writing.

    Related tools