Document Tampering Detection API

TrueDoc's fake document detection API runs forensic, structural, and AI-generation checks on every submission so your stack catches forged IDs, tampered PDFs, edited bank statements, fake invoices, and synthetic receipts in a single call.

Each verdict comes with per-field evidence — not just a score — so reviewers and audit teams can see exactly which fraud signals fired and where on the document.

The AI-generated document detection API layer evaluates latent signatures from current generative models, layout regularities common to generated PDFs and images, and rendering artifacts that betray synthetic origin.

The tampered document detection API layer inspects PDF object graphs, incremental updates, font dictionaries, ELA, splice and copy-move artifacts, and metadata edit trails — surfacing tampering even when the rendered document looks clean.

Both layers feed the same trust score and findings list, so your decisioning code never has to branch on signal type.

For async workflows, the document fraud detection API delivers signed webhook callbacks the moment a verdict is ready — no polling required. Each payload includes the verdict ID, trust score, top findings, and a link to the full evidence record.

Common automations: auto-accept clean verdicts in your onboarding funnel, route mid-range verdicts into a reviewer queue with evidence pre-attached, escalate only the strongest forgery signatures to fraud ops, and trigger trust-tier changes on marketplaces when re-verification fails.

Each /verify call returns a ranked list of fraud signals across four layers: visual forensics (ELA, splice detection, copy-move), document structure (PDF object edits, incremental updates, font substitutions), semantic checks (math reconciliation, MRZ checksums, date and amount consistency), and AI-generation signatures (latent fingerprints from current generative models).

Signals carry a confidence and a region or field pointer so your fraud reviewers can jump straight to the suspicious area without re-inspecting the document.

Treat the trust score as a soft input, not a hard gate. Combine it with your own velocity, device, and behavioral signals before deciding to accept, review, or decline.

For high-volume marketplaces, auto-accept verdicts above your threshold, send mid-range verdicts to a reviewer queue with the evidence pre-attached, and only auto-decline on the strongest forgery signatures to keep false positives low.

Fintech and lending: verify paystubs, bank statements, and proof-of-income at onboarding and underwriting. Auto-accept clean documents and route flagged ones to a fraud queue with evidence pre-attached.

Property management and tenant screening: verify income and ID documents at application. Reduce delays from manual paystub review and catch synthetic income proofs before lease approval.

HR, employment verification, and staffing: catch fake paystubs, diplomas, and reference letters during hiring. Plug into ATS and background-check workflows via API or webhooks.

Insurance: review claim documents, receipts, and supporting evidence for tampering and AI-generation signals before payout.

Compliance and risk: combine document trust scores with KYC and AML signals for a single decisioning view. Immutable audit logs and PII masking support regulator review.

Vendor and invoice review: detect fake invoices, edited receipts, and reused supporting documents in AP and procurement workflows.

The document fraud detection API supports signed webhook callbacks so async workflows — overnight underwriting, batch screening, offline claims review — receive verdicts the moment analysis completes.

Each callback carries verdict ID, trust score, top fraud signals, and a link to the full evidence record. Combine with your queue to auto-accept clean documents, route mid-range verdicts to a reviewer, and escalate only the strongest forgery signatures.