n8n Document Converter Node

n8n community node for converting documents to JSON/text. Supports 15+ formats with AI-friendly output.

Table of Contents

• Supported Formats
• Features
• Installation
• Usage
• Output Examples
• Architecture
• Development
• License

Supported Formats

Features

Core

• Automatic file type detection via magic bytes
• Strategy pattern: each format has its own processing pipeline
• DOCX output: plain text (default), HTML, or Markdown (GFM tables, headings, bold/italic)
• DOCX → Markdown ideal for AI/LLM/RAG pipelines
• XLSX multi-sheet processing with Excel-style column names (A, B, C...)
• JSON flattening for nested structures
• YML (Yandex Market) specialized parser
• usableAsTool: true for n8n AI Agent integration

Reliability

• Concurrency control via promise pool (Set-based, no race conditions)
• Fallback chains: DOCX uses officeparser -> mammoth, DOC/PPT uses CFB signature check + officeparser
• File name sanitization (path traversal protection)
• Configurable file size limits (up to 100MB)
• Custom error classes with descriptive messages

Installation

Via n8n UI (recommended)

Settings -> Community nodes -> Install
Package name: @mazix/n8n-nodes-converter-documents

Via CLI

cd ~/.n8n
npm install @mazix/n8n-nodes-converter-documents
# Restart n8n

Manual

git clone https://github.com/mazixs/n8n-node-converter-documents.git
cd n8n-node-converter-documents
npm install
npm run build
# Copy dist/ and package.json to ~/.n8n/custom-nodes/n8n-node-converter-documents/

Usage

1. Add "Convert File to JSON" node to your workflow
2. Connect a node that provides binary data (e.g., Read Binary File, HTTP Request)
3. Configure parameters:

Output Examples

Text document (DOCX, PDF, TXT, etc.)

{
  "text": "Extracted document content...",
  "metadata": {
    "fileName": "report.docx",
    "fileSize": 12345,
    "fileType": "docx",
    "processedAt": "2026-02-08T00:00:00.000Z"
  }
}

DOCX with HTML output

{
  "text": "<p>Introduction</p><table><tr><td>Name</td><td>Value</td></tr>...</table>",
  "metadata": { "fileName": "data.docx", "fileType": "docx" }
}

DOCX with Markdown output

{
  "text": "# Introduction\n\n| Name | Value |\n| --- | --- |\n| Item | 100 |\n\n**Bold text** and _italic_",
  "metadata": { "fileName": "data.docx", "fileType": "docx" }
}

XLSX (multi-sheet)

{
  "sheets": {
    "Products": [
      { "A": "ID", "B": "Name", "C": "Price" },
      { "A": 1, "B": "Apple", "C": 100 }
    ],
    "Orders": [
      { "A": "Order", "B": "Qty" },
      { "A": 101, "B": 5 }
    ]
  },
  "metadata": { "fileName": "data.xlsx", "fileType": "xlsx" }
}

JSON (flattened)

{
  "text": "{\n  \"user.name\": \"John\",\n  \"user.address.city\": \"London\"\n}",
  "warning": "Multi-level JSON structure was converted to flat object"
}

YML (Yandex Market)

{
  "text": "{ \"shop\": { \"name\": \"MyShop\" }, \"currencies\": [...], \"categories\": [...], \"offers\": [...] }"
}

Architecture

Project Structure

src/
├── FileToJsonNode.node.ts   # Node class (~220 lines)
├── types.ts                 # Interfaces (JsonResult, StrategyFn, YML types)
├── errors.ts                # Custom error classes
├── helpers.ts               # extractViaOfficeParser, limitExcelSheet
├── strategies/
│   └── index.ts             # All format strategies
├── processors/
│   └── yml.ts               # Yandex Market YML processor
└── utils/
    ├── sanitize.ts          # File name sanitization
    ├── promisePool.ts       # Concurrency control (Set-based)
    ├── columns.ts           # numberToColumn (1→A, 27→AA)
    ├── flatten.ts           # JSON flattening
    └── index.ts             # Barrel export

Processing Flow

Input binary → detect file type (magic bytes) → select strategy → process → output JSON
                                                      │
                              ┌────────────────────────┼────────────────────┐
                              │                        │                    │
                         Text formats            Spreadsheets         Special
                     (DOCX, PDF, TXT,          (XLSX, CSV, ODS)    (XML, JSON,
                      PPTX, HTML, ODT,                               YML, HTML)
                      ODP, DOC, PPT)

Technology Stack

Development

npm install          # Install dependencies
npm run build        # Compile TypeScript
npm test             # Run 57 tests (13 suites)
npm run lint         # ESLint check
npm run dev          # Watch mode
npm run test:coverage # Coverage report

CI/CD

• CI (ci.yml): lint -> build -> test -> security audit on push/PR to main/develop
• Auto Release (auto-release.yml): creates GitHub Release + git tag on version bump in main
• npm publish: manual (npm publish --access public)

Limitations

License

MIT © mazix

Links

• npm
• GitHub
• Changelog