Convert PDF bank statements to JSON

1. Add your PDF statements. Drag one or more PDF bank or credit card statements into the converter, or click to choose them. Password-protected files ask for the password, which is used only on your device.
2. Check the preview. Transactions appear in an editable table with dates, descriptions, money in, money out and balance. The balance check flags any row that does not add up, so you can fix it in a click.
3. Download JSON. Export to Excel (.xlsx), CSV, OFX or JSON, or copy the table straight into Google Sheets. Nothing is uploaded at any point.

JSON structure

Each export is a single object. Here is a shortened example:

{
  "generator": "bank-statements-converter.online",
  "generatedAt": "2026-11-03T10:15:00.000Z",
  "statements": [{ "file": "october.pdf", "pages": 1, "currency": "GBP",
                   "dateOrder": "DMY", "openingBalance": 2500, "closingBalance": 4061.23,
                   "balanceCheck": { "checked": 14, "passed": 14, "totalsMatch": true } }],
  "count": 14,
  "transactions": [
    { "date": "2026-10-02", "rawDate": "02/10/2026",
      "description": "Salary ACME Widgets Pty Ltd", "reference": null,
      "debit": null, "credit": 3200, "amount": 3200, "balance": 5700 },
    { "date": "2026-10-03", "rawDate": "03/10/2026",
      "description": "Rent Payment October", "reference": null,
      "debit": 1450, "credit": null, "amount": -1450, "balance": 4250 }
  ]
}

Designed for code

• ISO 8601 dates (YYYY-MM-DD), with the original text kept in rawDate for auditing.
• Numeric amounts with consistent signs: amount is negative for money out.
• Nulls, not empty strings, for missing values, so type checks stay simple.
• Metadata including the balance-check result, so your pipeline can reject statements that do not reconcile.

Typical uses

Load the file into pandas with pd.json_normalize(data["transactions"]), insert it into PostgreSQL or SQLite, feed a categorisation model, or seed test data for a fintech app. If you prefer tabular data, the same preview exports to CSV and Excel.

Privacy for sensitive data

Because parsing happens in the browser, you can turn real statements into JSON without sending them to a third-party API. That keeps you clear of data-processing agreements for this step and keeps customer data on your own machine.

• Guide: choosing a file format
• Bank statement to Excel
• Bank statement to CSV
• Bank statement to QuickBooks
• Bank statement to Xero
• Credit card statement converter
• Banks by country

A transactions array where each item has date (ISO), rawDate, description, reference, debit, credit, amount and balance, plus metadata per statement such as page count, detected currency, date order, opening and closing balances and the balance-check result.

Not at the moment. The converter runs entirely in the browser so that statements are never uploaded. You can export JSON from the page and load it into your own application.

Numbers. Debit and credit are positive numbers or null, amount is signed (money out negative), and balance is a number or null when the statement has no running balance.