DuckDuckGoose

DuckDuckGoose is an enterprise AI platform that detects deepfakes in selfies, identity documents, video and voice inside identity verification and fraud workflows. It is built for identity verification providers, fraud and compliance teams, and industries such as traditional banking, neobanks, fintech, government, insurance and media, and it returns an explainable verdict rather than a single opaque score.

Its models are trained on adversarial datasets from real fraud attempts and retrained continuously against more than 40 image, video and voice generation model families, so detection does not break when a new face-swap or generation model ships. Drop-in SDKs and sub-second API responses slot into an existing KYC, document or liveness flow without rearchitecting it.

DuckDuckGoose's product line includes Phocus, a web platform for fraud analysts and compliance reviewers to check images, video and audio for manipulation with an exportable audit trail; Waver, which detects deepfake speech across 16+ languages in real time; and DeepDetector, an on-premise detector for images and video. Each detection returns heatmaps, confidence scores and attack classification instead of a single score.

DuckDuckGoose integrates through drop-in SDKs and an API, running as a parallel forensic layer beside existing checks, a deepfake gate before costlier checks, or a retrospective sweep of already-onboarded accounts. It is GDPR compliant, EU-hosted on AWS Frankfurt, with ISO 27001 and SOC 2 Type II certification in progress; pricing is not published, and buyers are directed to contact the team or book a demo.

Use cases

• Stopping synthetic faces and AI-generated IDs at onboarding before a human reviewer sees them
• Catching forged, tampered or AI-generated identity documents
• Detecting morphed passport and ID photos before issuance
• Producing court-ready forensic analysis of manipulated image, video and audio
• Spotting manipulated claim photos and synthetic policyholders before an insurance payout