What this tool does
The Bank Statement Converter turns a bank or credit-card statement PDF into a clean spreadsheet. Drop the file, pick a format — CSV, XLSX, QIF, or OFX — and the file downloads. No signup, no per-page meter, no subscription.
The two paths
For Chase and Bank of America statements, the entire pipeline runs in your browser. The PDF is parsed by pdf.js (the same library that powers Firefox's PDF viewer), and a small bank-specific regex parser extracts each transaction into a structured row. The PDF never uploads. You can verify this in your browser's DevTools — the Network tab will show no outbound request containing the file or the extracted text.
For other banks, the tool extracts the text in your browser, then sends only the text — not the PDF bytes — to our AI service, which parses it into transactions and returns the structured data. We don't store the extracted text. The AI path is free and included in this tool — there's no upsell to a paid tier.
Output formats
- CSV — RFC 4180 compliant, opens cleanly in Excel, Google Sheets, Numbers, LibreOffice, and every accounting tool that imports CSV.
- XLSX — Excel format with typed columns (numbers stay numbers, dates stay dates) and reasonable column widths.
- QIF — Quicken Interchange Format. Imports into Quicken Classic, QuickBooks (with the QIF import option enabled), and most personal-finance tools.
- OFX 2.x — Open Financial Exchange. The modern accounting-import format used by QuickBooks Online, GnuCash, KMyMoney, and most bank-feed integrations. Each transaction gets a deterministic FITID so re-imports deduplicate correctly.
Sign convention
Debits — money out: purchases, ATM withdrawals, fees, payments — are negative. Credits — money in: deposits, refunds, interest — are positive. This matches OFX's TRNAMT and is the convention every spreadsheet expects when you sum a column to get net cash flow. A separate type column carries the literal "debit" or "credit" string so QIF and OFX exports stay correct.
Supported banks (v1)
The client-side parser ships with two banks in v1 — Chase and Bank of America — picked because they're the two largest US retail banks by deposit volume. Six more banks are fingerprinted (Wells Fargo, Capital One, American Express, Discover, Barclays, HSBC) but currently route to AI parsing until their client-side parsers ship in v1.1.
If your bank isn't on either list, the AI fallback handles it. Drop the PDF, click "Try AI parsing," and the tool will work the same way.
What doesn't work
- Scanned image-only PDFs without an OCR layer. If the PDF is a photo of a statement (or a scan from an old printer) with no text layer, neither the client-side parser nor the AI fallback can read it reliably. OCR the file first with Adobe Acrobat or a free tool like Tesseract, then re-drop the OCR'd PDF.
- Multi-account combined statements with several account-number sections in one PDF. The current parser treats them as one combined list. You'll see all the transactions, but you'll have to filter by date or account number afterward. (Splitting on the per-account headers is on the v2 backlog.)
- Statements over ~28 KB of extracted text bump against the AI fallback's request-size limit. The error message will suggest splitting the PDF; do that and re-run.
Privacy, in concrete terms
The privacy claim isn't a marketing line — it's something you can audit. For Chase and BofA statements, open your browser's DevTools, switch to the Network tab, and drop a statement. You'll see pdf.js fetch its worker file once, then nothing else. No POST to a server. No XHR carrying the PDF bytes. No request body containing the extracted text. The parser runs in your tab and the data lives in your tab's memory until you close it.
For the AI path, the same DevTools check will show one POST to /api/llm/bank-statement-converter with a body containing the extracted text. The PDF bytes are not in that body. Our AI service processes the request, returns the structured transactions, and we don't persist the input text after the response. You can re-audit at any time — the contract is the same call shape every other AI tool on this site uses.
Why this tool exists
The dominant tools in this category (bankstatementconverter.com, convertbankstatement.io, pdftables.com, moneythumb.com) gate the export behind either a signup wall, a per-page meter that runs out fast, or a recurring subscription for what's fundamentally a one-shot job. The honest version of "convert this PDF" is: do the conversion, hand back the file, charge nothing, store nothing. That's what this tool does.