OCR

The OCR node extracts text from images and document scans using Tesseract (via Tess4J). The recognized text is stored so it can be indexed and searched.

Kind

ocr

Applies to

Image

Inputs

None

Output keys

ocr_text (String)

Requirements

Tesseract / Tess4J plus its tessdata language files on the worker. CPU-bound; storage for the tessdata models.

Persists to

asset_json_comp + asset_node_result ledger

Configuration

OCRNodeOptions:

Option Meaning

tessDataPath

Path to the Tesseract tessdata directory

language

Recognition language(s), e.g. eng, deu

Use Cases

  • Searchable scans — index text from screenshots, receipts and scanned documents.

  • Document workflows — surface embedded text for classification or routing.

  • Compliance — find images that contain particular text.