Forensic Document Verification
Know whether a document was altered — in about a minute
Payslips, bank statements, invoices and IDs are trivial to forge and slow to verify by eye. Tamper Check runs 130+ forensic checks on the file and returns a plain verdict — naming the exact field that changed, not just a risk score.
- 130+ forensic checks
- Verdict in ~1 minute
- $0.50 per document
- 190+ countries
Developed by
TamperCheck.ai
Delivered, deployed and supported in the Philippines by Plexus Technology Corporation.
Three families of forensic signal
A convincing forgery has to survive all three at once — the file structure, the pixels, and the arithmetic.
Structural
What the file is made of, underneath what it displays.
- PDF object, font table and cross-reference inspection
- Incremental-update detection — evidence the file was edited after it was issued
- Editing-tool fingerprints left behind by common PDF editors
- Producer-metadata audit against what the issuing organisation would actually use
Pixel-level
What the image itself gives away, invisible to the eye.
- Error Level Analysis — edited regions compress differently from their surroundings
- Compression and noise inconsistency across the page
- AI-generation and deepfake detection via spectral and noise signatures
- Works on phone photos and scans, not only clean digital PDFs
Logical
Whether the document agrees with itself.
- Running-balance verification across every line of a statement
- Subtotals, tax lines and totals recomputed and compared
- MRZ integrity and checksum validation on identity documents
- Machine-readable fields checked against the visible ones
An answer your reviewer can act on
Not an opaque score between 0 and 100 — a stated conclusion, and the region of the document that produced it.
No tampering indicators found across the structural, pixel-level and logical checks.
Alteration detected — reported with the specific region or field that changed.
Not enough signal to decide, typically on very low-quality captures. Flagged for a human, not guessed at.
One endpoint, every document type
It slots into the intake step you already have, rather than adding a queue beside it.
- 01
Submit
Post the file to the API as multipart form data. PDF, JPG, PNG, HEIC and scans are all accepted.
- 02
Analyse
130+ forensic checks run across the three families — structural, pixel-level and logical.
- 03
Verdict
A result comes back in about a minute: authentic, tampered or inconclusive, with the altered field identified.
- 04
Act
Route the outcome into your existing process — proceed, escalate to manual review, or decline.
A single API call
Cloud pay-as-you-go, or deployed on-premise inside your own boundary.
Request
POST https://api.tampercheck.ai/v1/verify
Authorization: Bearer <api-key>
Content-Type: multipart/form-dataPlexus handles the integration into your intake, underwriting or onboarding workflow, so the check happens where the document already arrives.
Specifications
- Pricing
- $0.50 per document, pay-as-you-go, plus reasoning tokens at cost
- Enterprise
- On-premise deployment at custom pricing
- Accepted inputs
- PDF, JPG, PNG, HEIC and scanned hard copies
- Coverage
- 100+ document types across 190+ countries
- Uptime
- 99.9% SLA
- Transport
- TLS 1.3
- Data handling
- Analysed in memory, never stored, never used for model training. GDPR-aligned.
Anywhere a document is treated as evidence
If a decision turns on a document someone handed you, it is worth knowing the document is real.
Lending & credit
Insurance claims
KYC & onboarding
HR & hiring
Tenant screening
Education admissions
Visa & immigration
Healthcare & social care
Trades & construction
Plexus integrates Tamper Check into your document intake
We scope the document types you actually receive, wire the check into the step where they arrive, and set the thresholds that decide what reaches a human.
Request a consultation