Oz Forensics is a biometric identity verification company. It helps banks, fintechs and other online services confirm that the person opening an account or signing in is a real, live human and the same person shown on their ID, which stops fraud using photos, masks, replayed video and deepfakes.
It is built for regulated businesses that onboard customers remotely and need that check to be independently tested and certified.
At A Glance
| Vendor | Oz Forensics |
|---|---|
| Category | Biometric identity verification |
| Identities covered | Human users (liveness and face matching) |
| Deployment | On-premises or Oz managed SaaS; SDKs for iOS, Android, Flutter and Web |
| Customers | Evocabank, Alvos, Payme, Eurasian Bank, Freedom Finance Bank |
| Market classification | Identity Verification |
What Oz Forensics Does
Remote onboarding depends on a camera, and cameras can be fooled. Attackers hold up printed photos or masks, replay recorded video, or bypass the camera entirely by injecting synthetic video through virtual cameras and emulators.
Oz Forensics answers this at two points in the flow: first that a live person is present and the video stream is genuine, then that the face belongs to the claimed identity. Both checks share one back end and one set of capture SDKs, so a single onboarding flow does both. Detection relies on more than 50 machine learning models, and the system handles authentications in 30 countries.
Products
Oz Liveness
Oz Liveness is for onboarding and login journeys where a business must be sure a real person is in front of the camera. It defends against presentation attacks, injection attacks and deepfakes.
- Analyses the full video sequence frame by frame to spot inconsistencies, rather than judging a single image
- Detects virtual cameras and emulators in the SDK before any video is recorded
- Uses a passive selfie check by default, with a reported capture time of about 0.7 seconds
- Is trained on millions of synthetic attacks, including LoRA and LivePortrait deepfakes
Oz Biometry
Oz Biometry handles face matching, for organisations that need to link a selfie to a verified identity or find a person in their own records.
- 1:1 matching of a selfie to the photo on an ID document, with a reported accuracy of 99.99%
- 1:N identification across large face databases, used for fraud detection and VIP recognition
- Independent testing through NIST FRVT
How It Works
| Component | What it does |
|---|---|
| Oz API | The central component of the system. It exposes a RESTful interface to liveness and face matching analyses and contains file storage, a database, licensing logic and the Oz BIO module that runs the neural network models. The Full API works asynchronously with storage and webhooks, while the Instant API is synchronous and stores no data. |
| Mobile SDKs | iOS (Swift) and Android (Kotlin and Java) SDKs embed the capture interface in native apps and record the liveness video. |
| Web SDK | A Web Plugin in the browser talks to a Web Adapter on the server, which passes requests to the Oz API. The plugin also collects browser and camera data used to detect virtual cameras. |
| Web Console | A web interface for browsing data held by the Oz API. It keeps no data of its own. |
| Deployment | The whole stack can run on-premises, so biometric data never leaves the customer’s infrastructure, or as Oz managed SaaS. Small and proof of concept installations use Docker with Docker Compose; high-load installations use Kubernetes with Helm charts and an external PostgreSQL database, with Redis and nginx as supporting components. |
Identities And Systems Covered
| Human users | People verified during onboarding and authentication |
|---|---|
| Identity documents | The photo on an ID document, matched 1:1 against a selfie |
| Face databases | An organisation’s own records, searched 1:N |
| Client platforms | Web browsers and iOS, Android and Flutter applications |
Company Background
| Founded | 2017 |
|---|---|
| Leadership | Artem Gerasimov, Founder and CEO |
| Offices | Dubai (United Arab Emirates), Almaty (Kazakhstan), Porto Salvo and Lisbon (Portugal) |
| Ownership | Part of Unico ID Group since 2024, following direct investment from Unico Group |
| Research | An internal technology and R&D team of more than 70 engineers working on liveness detection and face matching |
| Scale | More than 3,000,000 reported authentications per day |
| Certifications | ISO/IEC 30107-3 Levels 1, 2 and 3 tested by iBeta and BixeLab, CEN/TS 18099 for injection attack detection, ISO/IEC 19795-2 for Oz Biometry by BixeLab, SOC 2 Type 2, ISO/IEC 27001 and GDPR compliance |
Note: This overview is compiled from Oz Forensics’s publicly available information as of 3 November 2026. It is provided for information only. NHI Mgmt Group does not endorse any vendor, and nothing on this page represents the opinions of NHI Mgmt Group.