Editing a PDF takes minutes. Hiding the edit is much harder — the earlier versions are often still saved inside the file. Indagar recovers every one of them, lines them up, and puts the whole file through eight forensic layers. What changed, when, and what doesn't add up.
The evidence is already in the file. A forged document rarely fails one check — it fails several at once, from the raw byte structure of the PDF to the arithmetic on the page. That goes for AI-generated fakes too: a payslip invented in one shot still has to carry a real ABN, sums that reconcile and the signature of genuine payroll software. The layers don't care who — or what — made the forgery. Every file goes through all eight, and every result is reported: passed, failed, or skipped because something couldn't be read cleanly. Observable facts you can point to on the document itself. Never a model's opinion. Never a guess.
Every earlier save buried inside the PDF, recovered and compared side by side.
Payees, employers and ABNs counted across every save. A genuine document has one of each.
One ABN attached to two different employers has no innocent explanation.
Every ABN validated against the ATO's 11-digit algorithm. An altered digit rarely survives.
Live against the Australian Business Register: exists, active, and really that employer's number.
Every readable figure recomputed — payslips, bank statements, rates notices, tax returns.
Lookalike fonts, single retyped figures, digits sitting off the line — the fingerprints of an edit.
Payroll software leaves one signature in the file. A PDF editor leaves another.
One engine, every document: the arithmetic adapts to whatever is in front of it, and everything else applies to any PDF at all — payslips, bank statements, rates notices, income statements, tax returns.
Written to be handed on — to a credit team, an auditor, or a dispute. Every layer listed with its result: passed, failed or skipped. A risk score with its reasons stated in full, save-by-save comparison tables, exhibit pages with the changed content boxed on the document image, and the file's SHA-256 fingerprint, so there is never an argument about which document was examined.
Nothing to install, no templates to set up, no training run. Indagar works on the PDFs you already receive — payslips, bank statements, rates notices, income statements and tax returns.
Drag a client PDF into the browser. Analysis starts on the spot and takes seconds — not a queue, not an overnight batch.
Eight forensic layers run in seconds: every saved version recovered and read, identities and ABNs compared, every figure recomputed, the typography and producer software examined.
A risk score with every reason laid out in plain English, and a report you can attach to the file. Indagar tells you what it found — never what to conclude. The decision stays yours.
The entire analysis happens on your own device. The document is never uploaded, never stored and never seen by us — there is no server copy to leak, because there is no server copy. Two optional checks send only the bare minimum when you switch them on: AI field-reading sends a rendered image of page one, and the ABN lookup sends the eleven digits of the ABN. Nothing else ever leaves the machine.
NOTHING STORED · NOTHING SHARED · SHA-256 EVIDENCE TRAIL IN EVERY REPORT