When your Banca Transilvania statement arrives as a scan or a photo of the page rather than a clean export, the numbers on it are just pixels until they are extracted. A statement document that was scanned or photographed, rather than saved as a clean digital file, is read with optical character recognition. The engine identifies the account number, date range and transaction table automatically, so no manual region selection is needed. Every extracted figure is then verified against the printed running balance: if a smudge or crease causes a 5 to be read as a 6, the mismatch is flagged on that row instead of remaining hidden.
If a transaction description on your Banca Transilvania statement is long enough to wrap across two or three lines, it still needs to land in a single row once converted. When a transaction description is long enough to wrap across two or more lines on the statement, it often confuses tools that expect one row per transaction. Convert·Into detects these wrapped descriptions and reconstructs them into a single record, preserving the full narrative text that is needed for audit and categorisation. The split is mended before the transaction table is populated, so the export does not contain orphaned continuation lines and the debit and credit amounts stay aligned with their correct descriptions.
If the running balance on your Banca Transilvania statement does not add up row by row, that is exactly the kind of mismatch Convert·Into is built to catch rather than pass through silently. Before the file is exported, the closing balance shown on the statement page is compared against every row in the reconstructed ledger. The running total is recalculated line by line, and any row where the arithmetic does not tally with the printed figure is noted. The output is not rejected when a single row fails: that row is flagged for review, and the rest of the statement is delivered clean. You can then export the verified data to Excel, CSV or QBO, with columns mapped into the fields your accounting package expects, and set about investigating only the handful of rows that genuinely need attention.