n8n-nodes-liteparse

An n8n community node for LiteParse — fast, local PDF & document parsing with OCR, screenshots, and markdown output.

Features

• Parse Documents — Extract text from PDF, DOCX, XLSX, PPTX, and images
• OCR Support — Built-in Tesseract or plug in any HTTP OCR server (EasyOCR, PaddleOCR, custom)
• Multiple Output Formats — Markdown (with headings, tables, lists), JSON (with bounding boxes), or plain text
• Screenshot Generation — Render document pages as PNG/JPEG images
• Batch Processing — Parse multiple documents in a single workflow
• 100% Local — No cloud dependencies, everything runs on your machine

Installation

In n8n, go to Settings → Community Nodes and install:

n8n-nodes-liteparse

Or install via npm in your n8n installation directory:

npm install n8n-nodes-liteparse

Operations

Parse Document

Parse a document and extract its content in your preferred format.

Generate Screenshots

Render document pages as images.

Batch Parse

Parse multiple documents from binary inputs in a single operation. Uses the same parameters as Parse Document.

Usage Examples

Basic PDF to Markdown

1. Read a file with the Read Binary File node
2. Connect to LiteParse with operation "Parse Document"
3. Set Output Format to "Markdown"
4. The output contains the parsed text in json.text

OCR a Scanned PDF

1. Read the scanned PDF
2. Connect to LiteParse with:
   • Operation: Parse Document
   • Enable OCR: true
   • OCR Language: eng (or your language)
   • Output Format: Markdown

Extract Screenshots for LLM Vision

1. Read a PDF
2. Connect to LiteParse with operation "Generate Screenshots"
3. Set Target Pages to 1-5
4. Screenshots are available as binary data on each output item

Use External OCR Server

1. Run an OCR server (e.g. EasyOCR or PaddleOCR)
2. In LiteParse node:
   • Enable OCR: true
   • Use External OCR Server: true
   • OCR Server URL: http://localhost:8080/ocr

Supported Input Formats

OCR Server API

If using an external OCR server, it must implement:

POST /ocr
Content-Type: multipart/form-data

Fields:
  file     - Image binary (required)
  language - Language code (optional, default: "en")

Response:
{
  "results": [
    {
      "text": "recognized text",
      "bbox": [x1, y1, x2, y2],
      "confidence": 0.95
    }
  ]
}

License

Apache-2.0