How to Parse Purchase Orders to JSON with an API

Every procurement workflow eventually hits the same wall: a PDF purchase order arrives and someone has to key the data into an ERP, an AP system, or a spreadsheet. At low volume it's annoying. At scale it becomes a real operational cost and a source of transcription errors.

The Parse Purchase Order Parser API removes that step entirely. POST a PO document, receive a clean JSON object with every field extracted and the arithmetic already validated. No configuration, no training data, no templates required.

What Gets Extracted

A single API call returns:

• •PO meta: PO number, issue date, expected delivery date, currency, payment terms, incoterms, and notes
• •Vendor: Name, address, email, phone, and tax ID
• •Buyer: Name, address, email, and phone
• •Ship-to: Separate delivery address when it differs from buyer
• •Line items: Line number, SKU, description, quantity, unit, unit price, line total, and per-line tax rate
• •Financials: Subtotal, discount amount, tax amount, shipping cost, and grand total
• •Validation: total_mismatch flag and a line_arithmetic_errors array listing every line where qty × unit_price ≠ line_total

The validation object is computed server-side — no extra API call needed. If line 3 has a pricing discrepancy, it shows up in line_arithmetic_errors with the expected total, the stated total, and the delta.

Getting Started

Sign up at cparse.com/dashboard and grab your API key. New accounts include free credit with no credit card required.

The base URL is https://api.cparse.com/po/v1.

Extract a PO: Python

import requests

url = "https://api.cparse.com/po/v1/parse"
headers = {"X-API-Key": "YOUR_API_KEY"}

with open("purchase_order.pdf", "rb") as f:
    response = requests.post(url, files={"file": f}, headers=headers)

result = response.json()[0]
data = result["data"]

print(data["po_number"])           # "PO-2026-00187"
print(data["vendor"]["name"])      # "Global Parts Supply Co."
print(data["grand_total"])         # 1832.40

# Check validation
if data["validation"]["total_mismatch"]:
    print("Warning: grand total does not reconcile with line items")

for err in data["validation"]["line_arithmetic_errors"]:
    print(f"Line {err['line_number']}: expected {err['expected_total']}, got {err['stated_total']} (delta {err['delta']})")

# Print line items
for item in data.get("line_items", []):
    print(f"{item['sku']}  {item['description']}  qty={item['quantity']}  total={item['line_total']}")

Extract a PO: JavaScript

import FormData from 'form-data';
import fs from 'fs';
import axios from 'axios';

const form = new FormData();
form.append('file', fs.createReadStream('purchase_order.pdf'));

const response = await axios.post('https://api.cparse.com/po/v1/parse', form, {
  headers: {
    ...form.getHeaders(),
    'X-API-Key': 'YOUR_API_KEY',
  },
});

const [result] = response.data;
const { data } = result;

console.log(data.po_number);           // "PO-2026-00187"
console.log(data.vendor.name);         // "Global Parts Supply Co."
console.log(data.grand_total);         // 1832.40

if (data.validation.total_mismatch) {
  console.warn('Grand total does not reconcile');
}

data.validation.line_arithmetic_errors.forEach(err => {
  console.warn(`Line ${err.line_number}: expected ${err.expected_total}, stated ${err.stated_total}`);
});

The Validation Object

The validation object is what makes this API useful for automated workflows, not just data display. Human-keyed POs regularly contain arithmetic errors. A line with qty=10, unit_price=$49.99 and line_total=$450.00 looks fine at a glance — the correct value is $499.90.

{
  "validation": {
    "total_mismatch": false,
    "line_arithmetic_errors": [
      {
        "line_number": 2,
        "expected_total": 499.90,
        "stated_total": 450.00,
        "delta": -49.90
      }
    ]
  }
}

• •total_mismatch is true when grand_total ≠ subtotal + tax + shipping − discount.
• •line_arithmetic_errors lists every line where qty × unit_price ≠ line_total, with the delta so you can decide whether to flag it, raise a query with the supplier, or treat it as a rounding difference.

Use the delta to set your own tolerance. A $0.01 rounding difference is probably fine to accept automatically. A $49.90 gap warrants a supplier query before payment.

Supported Formats

The API accepts PDF, DOCX, JPEG, PNG, and ZIP or 7z archives containing a single document. For scanned or image-based POs, OCR is handled automatically on the MEGA plan.

Integration Patterns

ERP ingestion: Extract line items and push them directly to your purchase order matching workflow in SAP, Oracle, or any ERP that exposes a REST API.

AP automation: On receipt of a supplier invoice, cross-reference po_number and vendor.name from the parsed PO to confirm the invoice matches an approved order before payment.

Supplier portal: Accept PO uploads from buyers, parse them automatically, and create orders in your fulfilment system without manual data entry.

Audit trail: Store the structured JSON alongside the original PDF. The validation object gives auditors a machine-readable record of any discrepancies found at intake.

Next Steps

See the complete field reference and plan details in the Purchase Order Parser documentation. Get your API key from cparse.com/dashboard.

Arthur Sterling is the Lead Developer Advocate at Parse.