Loading this page.
Loading this page.
Upload a general ledger journal export. F20 surfaces patterns worth a second look -- round-sum entries, weekend or period-boundary postings, one person preparing and approving their own entry, and reversals posted only days after the original.
Free journal analyzer: find round numbers, weekend postings and reversals worth questioning before an auditor does. Upload the file this page names, review the result in the browser, and download CSV or Excel when offered.
A journal with a weekend posting. These are synthetic fixtures, processed in memory and never stored.
Your file is processed in memory and never stored. Nothing is written to disk, and nothing is retained after the response.
Every pattern here has an innocent explanation -- an accrual is round by construction, a batch job posts on a weekend, a small team cannot always segregate preparer from approver. Each flagged line is worded as 'Potential anomaly requiring review,' a fact about the entry, never a conclusion about the person who posted it.
Scans a journal export for patterns worth a second look — round-sum entries, weekend postings, one person preparing and approving their own entry, a fast reversal, unusual concentration in one account or preparer — and reports each as a count with its evidence, never a verdict. Every one of these has an innocent explanation, so nothing here is fraud detection; the fixed wording for any such finding is exactly "Potential anomaly requiring review."
Upload a journal export. Round-sum entries, weekend postings, self-approved entries and fast reversals are reported as counts with evidence, never a verdict. Nothing here is fraud detection. Findings read "Potential anomaly requiring review."
Open the page, drop the journal export, and read the result in the browser. Download CSV or Excel when the tool offers those exports.
Yes. There is no charge and no account required. Every tool on F20 FinanceOps runs without signing in. A run made without an account is not recorded.
No. Uploaded files are processed in memory and dropped when the response is sent. Nothing is written to disk. They are never stored.
Have a journal export. Weekend postings, round sums, self-approved entries and fast reversals should stay in the file — deleting them before upload hides the review.
Read counts with evidence, not verdicts. Check accrual reversals next if period-end journals are the remaining question. Findings read "Potential anomaly requiring review." They are not an accusation.
Manual journal review is a close and audit prep step. Flags are deterministic counts. They are not an accusation and not an AI explanation of intent.