Home/Compare

Comparisons

TamperCheck vs the alternatives

18 comparisons, grouped by what the vendor actually checks. Every page answers the same fifteen questions - AI-generated document detection, how precisely a finding is located, pricing, retention, and what it takes to get started - sourced to that vendor's own published material.

7 comparisons

Document forensics

The file is the unit of work. These are the closest comparisons: every one of them examines an uploaded document rather than the person who sent it.

9 comparisons

Identity verification

The person is the unit of work - capture, liveness and face match against a library of ID templates. Strong on the physical security features of an ID, and not built for a payslip.

2 comparisons

Document platforms

Extraction first, fraud detection as one module beside it. The trade they offer is consolidation; the trade they ask for is depth.

Across all 18 vendors

Who keeps the document

For an identity platform the stored record is the product, and for a fraud platform the retained corpus is what the network signals and the human reviewers run on. Neither can give it up. We never write the document to disk, so there is no retention window to negotiate. Every line below is the vendor's own published position.

Published document-retention and model-training positions, by vendor
VendorDocument retentionTraining on your documents
TamperCheck.aiZero - analysed in memory, never written to diskNever
VerifyPDFEncrypted at rest, auto-deleted when the retention period expiresNot addressed in published material
KoncileStored in the EU - AWS Paris, DR in Frankfurt. SOC 2 Type II, HDSNot addressed in published material
InscribeDocuments retained - their network signals compare uploads across customersCross-customer network signals are central to the product
Resistant AINot addressed in published materialNot addressed in published material
EntrustVerification records retained for the audit trail; terms set by contractGoverned by contract, not published
JumioVerification records retained; extended retention is a priced add-onGoverned by contract, not published
Klippa DocHorizonEU-hosted, ISO 27001; documents held for the processing workflowCustom document types are trained on request
SnapptDocuments held in the portal for agent reviewModels trained on 16M+ submitted documents (their claim)
OcrolusDocuments retained - human review and analytics require itHuman-in-the-loop review feeds their models
PersonaVerification records retained; retention is configurableGoverned by contract, not published
MicroblinkOn-device capture limits what leaves the phone; server terms by contractTheir Fraud Lab trains models against emerging attack patterns
SumsubApplicant records retained for the compliance audit trailGoverned by contract, not published
RegulaCan run fully on-premise - you hold the dataOn-premise deployment keeps documents with you
MitekGoverned by contract, not publishedGoverned by contract, not published
FortiroNot stored beyond the initial processing period (their claim); ISO 27001Not addressed in published material
VeriffRecords retained; two-year extended retention costs +$0.30 per verificationGoverned by contract, not published
HyperVergeSet by contract and local regulationGoverned by contract, not published
FraudFinderClient Data held up to six years unless deletion is requestedTerms permit anonymised, aggregated Client Data for model development

Each vendor's position is taken from their own published material and dated on that vendor's comparison page. Two are worth reading twice: Veriff prices two-year retention as a +$0.30 per verification add-on, and FraudFinder's published terms allow Client Data retention of up to six years and permit anonymised, aggregated Client Data to be used in developing their models.

Still comparing?

Run your own documents through it instead. Add $5, we'll match it, and you can see a verdict on a file you already have.