LLMWhisperer icon

LLMWhisperer

Extract text from PDFs, images, and scanned documents using OCR. Preserves layout and formatting for accurate text extraction from any document type

Document → Extract Text

AI-generated

Overview

This node extracts text from PDFs, images, and scanned documents using OCR via the LLMWhisperer API. It supports preserving the layout and formatting of the original document for accurate text extraction. Common use cases include digitizing scanned documents, extracting text from images or PDFs for further processing, and preserving document structure for analysis or display.

Use Case Examples

  1. Extract text from a scanned PDF while preserving the original layout for accurate data extraction.
  2. Convert images containing text into editable text with options for high quality or low cost processing modes.
  3. Extract native text from documents quickly using the native text mode.

Properties

Name Meaning
File Contents The binary file contents to be processed for text extraction.
LLMWhisperer Host The host URL for the LLMWhisperer API endpoint.
Mode The mode to use for text extraction, affecting quality and cost.
Output Mode The output format of the extracted text.
Additional Options Additional optional parameters to customize the extraction process, such as adding line numbers, specifying page ranges, marking lines, and timeout settings.

Output

JSON

  • whisper_hash - Unique identifier for the text extraction job.
  • status - Current status of the extraction job (e.g., processing, processed, error).
  • text - Extracted text content from the document.
  • metadata - Metadata about the extraction process (removed from final output).

Dependencies

  • LLMWhisperer API with authentication token

Troubleshooting

  • Ensure the input binary data property specified in 'File Contents' exists and contains valid data.
  • Check the LLMWhisperer Host URL is correct and accessible.
  • Timeout errors may occur if the document is large or processing is slow; increase the timeout setting if needed.
  • API errors will return messages indicating the cause; verify API credentials and parameters if errors occur.

Links

Discussion