Step 1
Define the workflow and the rules
Choose the checks, document types and stop lists you need — different ones for new customers, large amounts or regions. Configured in your account, with no development.
An anti-fraud system that protects against fraud during customer verification
Fraud protection at the front door: NeuroVision checks the document, the face, the phone number and the sign-up history, and decides in seconds whether to let the customer in
- Fraud down to 0.01%
- 40+ anti-fraud document checks
- Data stays in Russia
Six fraud signals in one anti-fraud platform
Each module works on its own or inside a shared workflow — as in NeuroVision KYC. Switch on the ones your business needs.
Forged and edited documents
40+ anti-fraud document checks: the presence of mandatory security features, logical inconsistencies in the data, physical tampering and pasted-in elements, traces of a graphics editor, AI generation, a missing signature or stamp, wear and tear. 16,000+ document types from 195+ countries and territories, 90+ languages.
Someone else's face, a mask or a deepfake
We match the selfie against the document photo in 0.1 seconds with 99.74% accuracy — an algorithm ranked top-1 in Russia and top-24 worldwide by NIST FRVT. Liveness confirms that a live person is in front of the camera, not a photo, a video, a mask or a deepfake.
Repeat sign-ups and multi-accounting
A 1:N search across the face database: we recognise a person who has already signed up under a different name, with a different document or device. One person — one account; a blocked user will not come back under a new name.
Screen photos, printouts and replayed video
We tell a live capture of a document and a face from a screen photo, a photocopy, a printout and a recorded video.
Matches against blacklists and databases
We check the customer directly against state registers (the Ministry of Internal Affairs, the Federal Tax Service, the Federal Bailiffs Service, the traffic police, Rosfinmonitoring), against sanctions lists and 1,700+ AML databases, and against your own stop lists. Plus phone checks: whether the number belongs to the person, whether it is active and how long it has existed.
Anti-fraud scoring and rules
All the signals come together in a single decision according to your workflow: approve automatically, reject or send for manual review. You configure workflows, document sets and stop lists yourself in your account.
How the NeuroVision anti-fraud system works
Four steps from setting up the workflow to a decision on the customer. No development needed — everything is configured in your account.
Step 2
The customer goes through onboarding
The document and the selfie are captured in a Web SDK branded as yours; the full check — document, face, liveness — takes under 1.5 seconds.
Step 3
The platform gathers the signals and scores the risk
The document, the face, liveness, the phone number, the databases and the sign-up history are all checked in a single API call.
Step 4
You get the decision.
Approve, reject or send for manual review — through the API and in the dashboard. Every check is stored with a timestamp and a breakdown by source — an evidence trail for disputes and for the regulator.
Which fraud schemes it protects against
Five common schemes businesses meet during onboarding — and what NeuroVision uses to detect each one.
Signing up with someone else's documents
A fraudster submits a genuine passport that belongs to someone else
Matching the face against the document photo (99.74% accuracy) together with liveness makes it impossible to pass the check with someone else's document.
Generated and photoshopped documents
Documents assembled in a graphics editor or by a neural network
40+ anti-fraud document checks find traces of editing and generation; a passport's validity is confirmed with the Ministry of Internal Affairs and a driving licence with the traffic police.
Deepfakes and attacks on the video stream
A swapped face in the video, a screen photo or a recorded video instead of a live customer.
Liveness and replay detection tell a live person in front of the camera from any recording. More in the article «What a deepfake is».
Money mules and nominees
Nominee customers in whose name accounts and loans are opened.
Checks against Russian databases, stop lists and the phone number, plus a 1:N search for links between applications; for crypto services, transaction monitoring in KYT follows.
Multi-accounting and bonus hunting
One person signs up dozens of times to collect bonuses and promo codes.
A 1:N face search: one person — one account, even with a new document and a new device.
5
common onboarding fraud schemes
Anti-fraud for banks, fintech and online services
One platform for different industries — the verification workflows are tuned to each industry's needs.
-
Banks
- Anti-fraud in remote account opening and lending
- Protection against money mules
- 115-FZ, 152-FZ and 572-FZ — works without the Unified Biometric System
-
Microfinance and fintech
- Loans taken out on someone else's documents
- Repeat applications under different names
- A decision in seconds with no operator
-
Retail and marketplaces
- Seller verification under 289-FZ
- Fake orders and returns
- Multi-accounting for bonuses
-
The crypto industry
- Fake accounts, money mules and nominees
- AML screening against 1,700+ databases
- Paired with KYT
-
Car sharing and taxis
- Renting on someone else's licence
- Driver substitution — a selfie before unlocking, in 0.1 seconds
- Scheduled re-checks of driving licences with the traffic police
-
Bookmakers and iGaming
- Multi-accounting and bonus abuse
- 18+ verified by the document
- Self-exclusion lists checked before play is allowed
Why a business needs NeuroVision anti-fraud
Fewer fraudsters getting in, more honest customers getting through — and a minimum of manual work for compliance.
-
Fraudsters are stopped before the first transaction
Forged documents, other people's faces and deepfakes do not get through onboarding — before an account is opened or a loan is issued.
-
Honest customers are not put through extra checks
One simple action — placing the face inside the oval on the screen; the whole check takes under 1.5 seconds. Extra steps appear only when the risk is high.
-
Manual review only for borderline cases
Compliance sees the reasons for the risk on a single card; checks that could not be run are marked «not performed», not «passed».
-
Compliance with regulatory requirements
115-FZ, 152-FZ, 572-FZ; operation without the Unified Biometric System; data processed in Russia, on-premises inside the client's perimeter.
What the anti-fraud service costs
Flexible pricing with no subscription fee: you pay only for the checks that were performed.
You pay only for the checks performed
From 15 ₽ per document
Each additional document in the same session — from +8 ₽.
Checks against Russian databases
A package of 16 passport checks for 50 ₽.
Discounts as the volume grows
As volume grows, the price of document verification falls proportionally.
Modules are enabled individually
take only what you need
- Documents
- Face
- Liveness
- Databases
- Phone
- 1:N search
Why businesses choose NeuroVision anti-fraud
Modern business demands instant decisions, accuracy and security. Customers are not willing to wait, and competition does not forgive mistakes. Only technology that works fast, reliably and flawlessly gives companies an advantage.
NeuroVision — an expert in computer vision and AI
20 years in software development, 10+ years in AI and computer vision
Our own algorithms for document recognition, face recognition, liveness and anti-fraud — with no third-party engines. More than 10 billion faces and documents recognised.
Accuracy confirmed independently
The face verification algorithm is top-1 in Russia and top-24 worldwide by NIST FRVT 03/2023; 99.74% for faces, 99.85% for documents.
A Russian developer
NeuroVision LLC, a Skolkovo resident, listed in the Russian software registry (entry no. 19972 of 13.11.2023), holder of small technology company status (no. 5876), listed in the IT company register.
Live in one day
Integration takes under an hour and going live takes one working day. The trial period is a month on real traffic.
Terms
A 1-month trial on real traffic, a dedicated account manager, 24/7 support and payment only for the checks performed.
Licences and certificates
The project «Development of a multimodal system using artificial intelligence "Neuromed" for managing the diagnostic processes of a clinic on the basis of the "Neurovision" computer vision platform» is supported by the Innovation Promotion Foundation within the federal project «Domestic Solutions»
The Neurovision computer vision platform is registered with Rospatent (certificate No. 2023661508) and included in the Russian Software Registry (entry No. 19972 of 13.11.2023)
The right to use the Platform is granted under a licence agreement concluded by Neurovision with its customers
Frequent questions about NeuroVision anti-fraud
Briefly on what is asked most often: how it differs from KYC, setting up rules, pricing, timelines and data storage.
An anti-fraud system is a set of checks that identifies fraudsters before they become customers. NeuroVision checks the document, the face, the phone number, the sign-up history and the databases in a single API call and returns a decision according to your workflow: approve, reject or send for manual review. More in the article «Anti-fraud: what it is».
KYC is about establishing the customer's identity: recognising the document and matching it with the face. Anti-fraud checks that this identity has not been forged or reused: document authenticity, deepfakes, multi-accounting, stop lists. KYT is the monitoring of transactions as they happen. All three run on the same NeuroVision platform and can be combined into one workflow.
Forged and edited documents (40+ checks), other people's faces and deepfakes, screen photos and recorded video instead of a live person, repeat sign-ups by the same person under different names, and matches against sanctions lists, state registers and your own stop lists.
In your account you define the verification workflows, document sets, stop lists and the action taken on the result — automatic approval, rejection or manual review. Workflows can differ by product, amount and region.
There is no subscription fee — you pay only for the checks performed: from 15 ₽ per document, and from +8 ₽ for each further document in the same session. The trial period is one month on real traffic, with a dedicated account manager.
Integration through the REST API or the Web SDK takes under an hour on average, and going live takes one working day. There are SDKs for iOS and Android and a WebView option.
Data is processed in Russia: in the NeuroVision cloud or on-premises inside the client's perimeter. The solution complies with 152-FZ and 572-FZ and works without a connection to the Unified Biometric System. NeuroVision is listed in the register of personal data operators.
The customer photographs the document and takes a selfie in a widget branded as yours: for liveness it is enough to place the face inside the oval on the screen. The whole check takes under 1.5 seconds and requires no extra actions.
Still have questions? Ask us!
Contact us directly or send a request — we will reply within one working day
Can't put your request into words? Book a call and our team will contact you at a convenient time
Book a call
The company is included in the register of IT companies and the register of personal data operators of the Russian Ministry of Digital Development.
Address: Moscow, Presnenskaya Embankment 12 — Moscow City business centre, Federation Tower. Nearest metro stations: Delovoy Tsentr and Mezhdunarodnaya.
