BankStatementMateConvert a statement
BankStatementMate / Bank statement extractor
FROM STATEMENT TABLE TO TRANSACTION ROWS

Bank Statement Extractor for Transaction Tables

Extract supported bank statement transactions into five reviewable columns: Date, Description, Debit, Credit and Balance. Use the preview to compare the result with the source, then download Excel or CSV.

Extract statement transactions ↗
Focused extraction, not a general document parser

BankStatementMate is built around transaction tables rather than arbitrary PDF content. For supported layouts it separates dates, descriptions, money out, money in and running balances so the result can be reviewed as a spreadsheet table.

What the bank statement extractor returns

The current output contains five transaction fields: Date, Description, Debit, Credit and Balance. Excel exports use numeric amount cells; CSV exports contain the same transaction columns in a portable text format. Dates are normalized only when their order is clear. A source date without a year can be preserved with a review notice rather than silently inventing a year.

How statement extraction works

  1. Open the statement. Digital PDFs use positioned text extraction. Clear scans and images can use local OCR.
  2. Find the transaction table. The parser looks for supported headings and uses their positions to assign values to the correct columns.
  3. Keep transaction order. Rows remain in statement order so running balances can be checked before you sort or edit the data elsewhere.
  4. Preview the extracted rows. The browser shows the transactions before a download is offered.
  5. Export. Download Excel for spreadsheet review or CSV for a portable table.

Checks that help prevent silent bad exports

The extractor is deliberately conservative. Missing or ambiguous monetary columns can stop the export instead of being guessed. Running-balance mismatches are surfaced for review. Explicitly labelled transaction-detail sections can be reconciled against a main balance table when account, date, direction, amount and description evidence agree; unmatched detail entries block the full export rather than being silently dropped.

These checks reduce one class of extraction error, but they do not certify the statement. Compare the output with the original before bookkeeping, accounting or financial use.

Current support and boundaries

  • PDF, PNG and JPG input; 10 MB maximum per file and five PDF pages.
  • Clear English transaction tables with separate Date, Description, Debit, Credit and Balance headings are the current target.
  • Single signed Amount columns, unusual wrapped layouts or unclear scans may be rejected.
  • The interface processes one file at a time.
  • The tool does not categorize spending, connect to a bank account or decide how transactions belong in your books.
  • The statement file is processed in the browser; account APIs store account and allowance data, not statement files or extracted transactions.

Follow the PDF transaction extraction walkthrough for the next step.

After extraction, labels require a separate workflow. See how to categorize bank statement transactions with rules, AI suggestions and review. BankStatementMate exports five fields; it does not generate transaction categories.