Fight fraud with the strength of the Stripe network

Radar evaluates transactions in real-time, using machine learning algorithms to assess the risk of fraud based on hundreds of data points. For this reason, both tiers of Radar charge a fee per screened transaction, meaning any payment attempt that Radar evaluates, regardless of outcome. This includes:

• Successful charges
• Declined transactions
• Blocked payments
• Payments flagged for review

Use ML that actually works

Radar detects and blocks fraud using machine learning that trains on data across millions of global companies.

Stop blocking legitimate customers

Old ways of combating fraud were never designed for modern internet businesses and can lead to lower acceptance rates and lost revenue. Radar helps you distinguish fraudsters from customers and apply Dynamic 3D Secure to high-risk payments.

Get protection that adapts to you

Our machine learning infrastructure lets us retrain models – including ones customised for your business – every day. Radar’s algorithms quickly adapt to shifting fraud patterns and to your unique business.

No integration required

Radar is built into Stripe – get started from day one with zero code. And it’s seamlessly integrated with other products, like Checkout, so you can access the full Stripe platform.

Trained with hundreds of billions of data points

Stripe processes payments from 197 countries for every industry, company size, and business model. Even if a card is new to your business, there’s a 91% chance it’s been seen before on the Stripe network.

Strength in numbers

By learning from millions of global businesses processing billions in payments each year, Radar can assign risk scores to every payment and automatically block many high-risk payments.

Better ML outcomes with Stripe-scale data

Radar scans every payment using thousands of signals from across the Stripe network to help detect and prevent fraud – even before it hits your business.

More accurate than third-party tools

Radar ingests data from every layer of the financial stack, integrating checkout flow data, rich payments data, and card network and bank information to assess fraud risk.

Data from financial partners

Stripe’s partnerships with Visa, Mastercard, American Express, and leading banks let us use data like TC40s, SAFE reports, and early dispute notifications to help identify fraudulent charges before they’re disputed.

Data from payments

Radar is fully integrated with the rest of Stripe, so our systems can instantly use rich payment information – such as customer details, shipping and billing addresses, and other properties – to improve machine learning performance.

Data from checkout flows

Stripe’s checkout tools automatically incorporate certain buyer patterns from your website and mobile app to help detect anomalous payments. Radar uses these signals to map a comprehensive customer path that can accurately predict fraud.

Sophisticated signals engineered for impact

Radar is built for global internet businesses facing evolving fraud threats. Radar constantly tweaks our algorithms, tests which attributes are most relevant, and generates compound signals to help precisely identify and block fraud.

Device fingerprints

Tie multiple signals into a single device profile.

Historical snapshots

Spot recurring patterns across years of historical data.

Proxy detection

Identify IP spoofing and proxy usage.

Cumulative impact

Radar scores every transaction by combining multiple signals to help determine whether a payment is fraudulent.