My bank only offers PDF statements.

Drop the PDF into the converter — standard tabular layouts parse well and the parser reports its confidence. Scanned or image-only PDFs are detected and rejected with guidance, because there is no text to read; ask the bank for a digital original or a CSV.

My CSV uses semicolons and amounts like 1.500,00.

That is the continental-European convention, and it is auto-detected file by file — the delimiter is sniffed from the rows and decimal-comma amounts parse correctly once the file proves the convention.

The column headers are in another language or unusual.

Common header variants are matched automatically. When they are not recognized, a manual column mapper appears so you can point at the date, description, and amount columns yourself.

Get a CSV out of any bank.

Every online banking portal can produce a CSV — the trick is knowing where to look and which format to prefer. This guide covers the patterns that fit almost every bank, and what to do with the file afterwards.

Why CSV is the export to ask for

A CSV is structured data: every transaction is a row, every field is a column, and nothing depends on how the bank chose to lay out a page. PDF statements are pictures of tables by comparison — parseable when the layout is standard, fragile when it is not. If you plan to analyze or convert a statement, CSV (or OFX) gives the cleanest result every time.

Where banks hide the export

Log into online banking and look for a section called Statements, Documents, Transactions, or Transaction history — the wording varies, the pattern does not. Open the account, set a date range, and look for a Download or Export button; it usually offers a choice of formats. Mobile apps often bury the same feature under the account’s detail screen or a share icon.

Which format to pick when the bank offers several

If the destination is accounting software, prefer OFX — it is the format those tools import natively. For spreadsheets and general analysis, CSV or Excel. Treat PDF as the fallback for when nothing structured is offered: standard tabular PDFs parse well, but multi-line layouts can parse partially, and scanned or photographed statements do not parse at all.

Set the date range deliberately

Export whole months, and go back as far as the analysis needs — many banks cap the online range at 12 to 24 months, so older history may need a support request. If your exports overlap, the PesaInsights dashboard merges duplicates so the same transaction is only counted once.

What to do with the CSV

Drop it into the converter for Excel, OFX, QIF, IIF, or JSON output, or into the analyzer for a spending dashboard — both run in your browser. Delimiters (comma, semicolon, tab, pipe) and European decimal-comma amounts are detected per file, and when the column headers are unusual a manual mapper lets you point at the date, description, and amount columns yourself.

Convert a statement · Analyze a statement · Help & FAQ · Privacy promise