Invoice & Receipt Parser API: PDF/Image to JSON

Use when: a bookkeeping, expense or AP agent needs supplier, invoice number, dates, currency, subtotal, VAT/tax, total and line items from an invoice or receipt (PDF, scan or photo) as JSON, without paying for an LLM call.

Try this input

{"requests": [{"url": "https://raw.githubusercontent.com/invoice-x/invoice2data/master/tests/compare/AmazonWebServices.pdf"}, {"text": "ACME Ltd\nInvoice No: 1042\nTotal £42.60"}]}

Sample output (one item, shortened)

{"method": "pdf_text", "supplier": "Amazon Web Services", "invoice_number": "42183017", "invoice_date": "2014-08-03",
 "currency": "USD", "tax": 0, "total": 4.11,
 "line_items": [{"description": "Amazon Glacier", "amount": 2.22}, "..."], "confidence": 1, "error": null}

Text PDFs are read directly (layout-aware); scanned PDFs and photos (PNG/JPG) go through Tesseract OCR. Also extracts VAT/tax IDs, IBANs and emails, plus an evidence line for totals. Up to 200 items per run.

Pricing: pay per event, $0.01 per document parsed. Rows with errors are not charged.

Limits & failure modes

• Rule-based, best on typed English-language invoices. Unusual layouts can miss fields (returned as null).
• OCR on photos is weaker: expect totals and currency, but supplier names and line items may be wrong or missing. Check confidence (0-1) and double-check below about 0.8.
• URLs must be public; private/internal addresses are refused. Bad or unreachable files return a row with error.
• Documents are processed in memory and not stored.

Last tested 2026-10-29.

Use from AI agents (MCP)

• Apify MCP server: add https://mcp.apify.com/?tools=tudorlane/invoice-receipt-parse to Claude, Cursor, ChatGPT, VS Code or any MCP client.
• REST: POST https://api.apify.com/v2/acts/tudorlane~invoice-receipt-parse/run-sync-get-dataset-items?token=YOUR_TOKEN with the input JSON.