OCR server supported languages

The Nutrient OCR component supports a wide range of languages, enabling precise text recognition based on linguistic characteristics such as ligatures, punctuation rules, and symbol variations. To ensure accurate text extraction, specify the language of the document during OCR configuration.

If your required language isn’t listed, contact Support for assistance.

Supported languages

Usage

OCR capabilities are exposed through the following API endpoints:

• OCR when uploading documents
• OCR to process and download a document
• OCR to edit and save a previously uploaded document

You can specify language of your document using either:

• Full language name e.g. english, german — available for commonly used languages
• ISO 639-2 language code e.g. eng, deu — available for all languages
• ISO 639-2 language code with variant e.g. chi_sim_vert or deu_frak

OCR a document with one language

Below is an example of a curl request to Document Engine to OCR a document in Japanese (ISO 639-2 code jpn):

# Assuming Document Engine is running on `localhost:5000`.

curl -X POST http://localhost:5000/api/build \
    -H 'Authorization: Token token=secret' \
    -o jpn-ocr-result.pdf \
    --fail \
    -H 'Content-Type: multipart/form-data' \
    -F 'scanned=@/path/to/japanese-document.png' \
    -F instructions='{
      "parts": [
        {
          "file": "scanned"
        }
      ],
      "actions": [
        {
          "type": "ocr",
          "language": "jpn"
        }
      ],
      "output": {
        "type": "pdf"
      }
    }'

OCR a document with multiple languages

To perform OCR on a document containing multiple languages, specify a list of desired languages (or their ISO 639-2 codes). Below is an example of a curl request to Document Engine to OCR a document with two languages, English and French:

# Assuming Document Engine is running on `localhost:5000`.

curl -X POST http://localhost:5000/api/build \
    -H 'Authorization: Token token=secret' \
    -o french-english-ocr-result.pdf \
    --fail \
    -H 'Content-Type: multipart/form-data' \
    -F 'scanned=@/path/to/english-french-document.png' \
    -F instructions='{
      "parts": [
        {
          "file": "scanned"
        }
      ],
      "actions": [
        {
          "type": "ocr",
          "language": ["french", "eng"]
        }
      ],
      "output": {
        "type": "pdf"
      }
    }'