The two ledgers don’t need to look alike. The bridge still needs to reconcile.

Your files are checked before anything is compared.

Berean detects which product wrote each export rather than asking you, then refuses anything it cannot read safely: a missing required column, a row with no account or no date, a property-software row with no amount, a file with no rows or more than it will accept, journals whose debits and credits do not balance, section totals that disagree with the rows under them, and two files on different accounting bases.

Some of those refusals name the exact problem — a column that is not there, a file with no rows, two files on different bases. Others report how many rows did not check out and refuse the file whole. Either way nothing is read from either export, and nothing is compared until you fix it.

When a file cannot be read at all, the message carries a short reference code. Quoting it tells Berean exactly which check stopped it, and the code travels without carrying a ledger row.

The free audit does not keep your ledger.

It runs entirely in memory. Nothing about a free-audit ledger is written to a database or to disk — the comparison runs, you read the result, and the rows are gone when the request ends. What comes back on screen is aggregates: counts, dollar totals, dates and per-check subtotals. It structurally cannot contain a payee, a memo or a ledger line.

Berean reads these exports. It never writes to either system. Your exports are read in memory, compared, and not retained — no ledger row from a free audit is written to a database or to disk. Upload only data your company is authorised to process: Berean does not assume an export is free of sensitive information, and it never writes back to either system.

Three fair objections.

Aren’t the two systems supposed to differ?

Yes — that is the point. Berean separates what matched transaction-to-transaction, what matched because detail rolled up into a monthly summary entry, and what is left unexplained. A month your bookkeeper summarises in one journal entry is matched to its detail rather than reported as dozens of missing transactions — when the two sides reconcile exactly. When they nearly do, or when Berean cannot tell which entry summarises which, it says so and reports the transactions individually rather than matching on a guess.

My bookkeeper already does this.

Keep them. Berean hands them a deterministic exceptions list with the source rows attached, so month-end starts at the evidence instead of starting at the search. The same two files always produce the same list.

Is exporting my data even allowed?

A general-ledger export is a standard feature of both platforms and you run it yourself — Berean never signs in anywhere on your behalf. Upload only data your company is authorised to process.