The OCR node extracts text from image-based documents or raw images using Optical Character Recognition (OCR). This is useful for converting scanned files, screenshots, or image-based documents into machine-readable text for downstream processing.
Inputs
- Documents – Accepts structured document-like files (e.g., PDFs or multi-page scans) as input for OCR processing.
- Image – Accepts raw image files (e.g., PNG, JPG) that contain embedded text to be recognized.
Output
- Text – Emits extracted plain text from either the “Documents” or “Image” input. The output can be used in embedding nodes, LLM prompts, or stored for further analysis
Configuration Steps
- Type of OCR – Select the OCR mode based on the language structure:
- Default – for general mono- or multilingual OCR processing.
- Language – Choose the language that best matches the input text.
- Example – English
