n8n-nodes-parsedoc

n8n community node for ParseDoc — turn any receipt or invoice (image or PDF) into structured, accounting-ready data, right inside your workflows.

Stop copying numbers off receipts by hand. Drop a ParseDoc node after any trigger that produces a file (email attachment, Drive upload, webhook, scanner) and get back a clean accounting line: date, merchant, total, tax, category, payment method.

What it does

• Parse to Accounting Line (default): one flat row per document — date, merchant, total, tax, currency, category, payment_method, needs_review — ready to push into Google Sheets, Airtable, a database, or your accounting tool.
• Parse Document: the full structured JSON (every line item, multiple tax rates, merchant address, invoice number, confidence, warnings).

Works on receipts and invoices in many languages and currencies. needs_review flags low-confidence extractions so you can route them to a human only when it matters.

Installation

In n8n: Settings → Community Nodes → Install → n8n-nodes-parsedoc.

Usage

1. Add a trigger that outputs a file (e.g. Gmail Trigger with attachment, Google Drive, Webhook).
2. Add the ParseDoc node.
3. Set Document Source to Binary Property (the attachment) or URL.
4. Pick Parse to Accounting Line and connect the output to your spreadsheet/DB node.

That's it — every time the workflow runs, the document is parsed automatically. No manual entry.

Credentials

The node works out of the box on a free demo tier (rate-limited, for testing).

For production volume, add ParseDoc API credentials:

• API Key — your key from ParseDoc (starts with pk_).
• Base URL — leave as default unless self-hosting.

Pricing is usage-based ($0.02/page). Failed parses are never billed.

Example output (Accounting Line)

{
  "date": "2026-06-09",
  "merchant": "Starbucks Coffee #1147",
  "currency": "USD",
  "subtotal": 11.95,
  "tax": 1.06,
  "total": 15.01,
  "category": "restaurant",
  "payment_method": "VISA",
  "item_count": 2,
  "confidence": 0.95,
  "needs_review": false
}

License

MIT