Cryptographic verification is now live. Govern, attest and mint every check.Go Now

deepidv × scam.ai

If it isn't real,
we'll catch it.

Frame-level detection of face swaps, presentation attacks, and AI-generated synthetic media across video, image, and voice. Powered by deepcam — the only deepfake stack with hardware and software in one platform.

Book a demo

DocumentationDeveloper docs

deepidv detects synthetic media, explains how it was made, and responds in real time — from the first frame to the final decision.

Vectors scanned

Face swapGAN faceMorphReplay

Catch every synthetic vector

Face swaps, GAN-generated faces, morphs, manipulated video, voice clones, and presentation attacks — flagged the moment they hit your flow.

Confidence0.98

Pixel

Geometry

Temporal

Understand exactly why

Pixel-level forensics, facial-geometry analysis, and temporal checks explain every verdict with a calibrated confidence score.

Decision<1s

RejectStep-up authManual review

Respond in milliseconds

Reject, step-up, or route for review through a single API call — before a fraudulent account is ever created.

How It Works

See deepfake detection in action

Watch how deepidv catches synthetic media in real time.

Comprehensive Detection

Every synthetic vector, caught

Generative tools evolve weekly. deepidv's models are trained on the full landscape of attacks — so face swaps, AI-generated faces, morphs, and presentation attacks collapse under forensic scrutiny.

Detection coverage

AI-Generated Faces

GAN & StyleGAN artifacts

Face Swap & Replacement

DeepFaceLab, FaceSwap

Morphed Photos

Blended hybrid identities

Manipulated Videos

Temporal & frame-level

Audio Deepfakes

Voice clones & synthesis

Presentation Attacks

3D masks, replays, screens

deepcam · forensic scananalyzing

Forensic view

As submitted

Synthetic media98% · face_swap

Detected vectorsAuto-rejected

face_swapgan_facemorphreplay

How It Works

Multi-layer forensic analysis

Every image and video is examined at four forensic levels simultaneously. No single signal decides — the layers corroborate, and the result is a confidence score you can branch your logic on.

Four forensic layers

Pixel-Level Forensics

Noise & compression artifacts

Facial Feature Analysis

Anatomical consistency

Temporal & Biological

Motion & micro-expressions

Source & Liveness

Tool signatures & living subject

deepcam · forensic enginecorroborating

Aggregate confidence

0.98/ 1.00

Synthetic

Pixel forensics96

Facial geometry92

Temporal motion88

Source & liveness70

4 / 4 signals corroborateno single point decides

Flexible Integration

Built into your workflow

Drop deepfake detection into onboarding, transaction authorization, or batch review. Each integration returns the same confidence score — feed it straight into your fraud rules engine.

Ways to integrate

Image Analysis API

Photos & selfies

Video Analysis API

Streams & recordings

Real-Time SDK

Mobile & web capture

Batch Processing

High-volume media

Read the docs

POST /v1/deepfake/scan200 ok

Inputface_capture_4821.jpg

→

{
"verdict": "deepfake",
"confidence": 0.98,
"vectors": ["face_swap"],
"action": "reject"
}

Image APIVideo APIReal-Time SDKBatch

Trusted & Certified

Built to meet the world's most demanding security and privacy standards.

SOC 2 Type II

Audited security controls for data protection

GDPR Compliant

Full compliance with EU data privacy regulations

ISO 27001

International standard for information security

DIACC Member

Digital ID & Authentication Council of Canada

Integration

Take full control.
Getting started takes minutes.

Plug deepidv into your stack three ways — a REST API, native SDKs, or an MCP server your AI agents can call directly. Clean docs, a sandbox, and one source of truth take you from API key to live verification the same day.

REST API

Call any verification service directly — create a session or run a standalone check and get structured JSON back in under 500ms.

→ screening 200+ watchlists…
✓ status: clear
  risk_score: 4 · 143ms

Drop-in SDK

Pre-built UI components and server libraries for iOS, Android, Web, Node, Python, and Go — zero to live verification in a single sprint.

  checks: ['id','liveness','face'],
✓ v.url → send to your user

MCP for agents

An open protocol that lets any AI agent call deepidv's verification engine directly — no custom integration code. Connect it in one command.

Claude

Launch deepidv for my AI agent — set it up from the docs.

✓ 12 verification tools available

Ready to run ID, liveness & AML checks.

Changelog

The Latest

See all releases

May 18, 2026Release

Verify 2.0 is here

Multi-language verification, a redesigned ID & face-capture flow, end-to-end whitelabel, and a native Background Check step.

May 9, 2026Features

ProofCall, smoother capture & translated reports

AI background-confirmation calls, faster camera start-up, translated PDF/HTML reports, and location-by-IP enrichment.

April 10, 2026AI

AI-powered adverse media screening

Continuous, AI-driven adverse-media checks built natively into your verification workflows.

See all releases

Frequently asked questions

deepidv detects a full spectrum of AI-generated and manipulated media: GAN/StyleGAN-generated faces, face swaps (DeepFaceLab, FaceSwap), morphed photos, manipulated videos, audio deepfakes and voice clones, and presentation attacks including 3D masks, silicone masks, and photo/video/screen replays.

Real-time detection returns results in under a second during live verification sessions, enabling immediate rejection of fraudulent accounts at the point of creation without adding perceptible friction for legitimate users.

Our multi-layer detection analysis achieves 90%+ accuracy by combining pixel-level forensics, facial feature analysis, temporal and biological analysis, and source/liveness verification. Confidence scores are returned with every analysis to support your risk decisioning.

Real-time detection analyzes images or videos during live verification and returns results in seconds, ideal for onboarding and transaction authorization. Forensic analysis performs deep examination of suspicious media and produces comprehensive reports used for investigations, reviews, and legal proceedings.

We offer an Image Analysis API, Video Analysis API, Real-Time SDK for mobile and web, and Batch Processing for large volumes of existing media. Each integration returns a confidence score you can feed directly into your fraud rules engine.

Still have questions?

Our team is ready to help you get started.

Contact Sales

All-In-One Products

Explore the Full Suite

Background Checks

Credit Checks

Deepfake Detection

Education Confirmation

Financial Screening