Bank statement PDF to Excel, then answers, not just rows

The problem

Your bank gives you PDFs. You need rows: date, description, amount, so you can see where money went or prove a deposit arrived. Copy-pasting from a PDF breaks columns, and older months are often PDF-only.

How it's solved today

Dedicated converters (DocuClipper, Nanonets and many others) turn statements into Excel, CSV or QuickBooks files. Some run in the browser; others process on their servers. Then you still build the analysis yourself in a spreadsheet.

How SandQuery helps

Reads the PDF, including scans and photos. Text PDFs and scanned pages are read, and the rows are checked against the statement's own numbers.
You review what it isn't sure of. Anything a check can't confirm goes to you to review line by line before it's used. If a file can't be read, the upload fails instead of producing plausible-looking rows.
Years of statements as one. Keep each month in a set; reports read the set as one table, and you add each new month as it arrives.
Then ask the question. Run reports on the rows: Duplicate payment detector, Subscriptions and recurring charges, Vendor concentration, Statement rollforward, Stripe payouts matched to bank deposits.
Download the rows and report results as CSV, which opens straight in Excel and Google Sheets.

Want native Excel files?

FAQ

Which banks are supported?

Any bank whose statements are in English or another language written in the Latin alphabet. Each PDF can be up to 20 MB and 100 pages.

Is the PDF sent to an AI model?

Yes. An extraction service on SandQuery's private network pulls out the text or page images, and a model reads them into rows. The models are self-hosted, with OpenRouter as a fallback. See Security.

Is it free?

Uploading a statement is free. Each report shows its price before you run it.

Next step

Join the waitlist · Find duplicate payments · Stripe payout reconciliation