Platform overview

Comprehensive tools to streamline your risk operations


Single interface for KYC/KYB, investigations, AML monitoring, and case management

Smart Signals

Billions of device, behavior, identity, financial and transaction data points

Rule Editor

No-code rule builder allows you to evaluate 4,000+ signals and custom data fields


Pre-built rulesets for identity verification, payments, counterparty risk, and AML

ML Models

Machine learning models trained across billions of sessions to detect fraud

Network Graph

Explore suspicious connections between users and devices in a visual interface

Anomaly Detection

Detect anomalous transactions in real-time and flag emerging “zero day” fraud threats


Detailed reporting on fraud KPIs, rule performance, transactions, and compliance

Data Enrichment

Integrated with 35+ providers A/B tested for data coverage, performance, and quality

“I've never used a tool that's this flexible for fraud fighters or worked with a customer support team that's so accessible and effective.”
How it works

Engineered to level up your fraud operations

Every part of our platform works together to detect and stop fraud with high precision. We capture data across the customer lifecycle and provide full transparency into risk decisioning.


One lightweight integration gives you access to device intelligence, behavior biometrics, real-time transaction data, and our growing network of tagged devices, users, and counterparties.

Smart Signals

Our software analyzes deep device-level signals and user behavior patterns during active sessions to identify signs of fraud. We cross-check this against our consortium and enrich it with identity, email, phone, banking and payment data to provide a 360° view of risk.

Feature Store

Using raw evaluation data from billions of sessions and transactions, we’ve engineered over 4,000 expert fraud features. These features can be evaluated in real-time using our rules engine and are used to train our machine learning models.

We can also make them available to your data science and machine learning teams for training custom ML risk models.

Machine Learning

We apply machine learning to accurately predict identity fraud, payment fraud, counterparty risk, and money laundering activities. Our models have been trained across billions of sessions and can output risk scores in 100s of msec.

Rules and Risk Scoring

Sardine comes bundled with hundreds of granular rules for fraud risk scoring. Our no-code rule editor allows you to evaluate over 4,000 features in real-time, assign weights for more precise risk scoring, and upload custom data fields to score your own metrics.

Rules can be back-tested against historical data or run in shadow mode to gauge performance using live transactions.

Decisioning and Feedback

When a rule is triggered, our workflow automatically approves, declines, or flags a transaction for manual review. Your risk decision then gets funneled back into our Feedback API to continuously improve our machine learning models.

Why choose Sardine

Our tools are customizable, easy-to-use, and designed to accurately and swiftly detect risky behavior across different use cases and industries.

Smart Device + Behavior signals in one SDK

Sardine uniquely combines device and behavior into one SDK — built by the same team that built the fraud detection systems at PayPal, Uber, and Coinbase. The alternative requires integrating with two separate providers and managing different dashboards, tools, data sources, and decisioning.

Consolidate dozens of vendors in one Platform

Sardine consolidates all the tools and data you need in one place to make your fraud operations more efficient. You don’t need to switch between dashboards during investigations. We also integrate 35+ vendors so you can do more without requiring additional engineering resources.

We act as an extension of your team

We partner closely with you to tackle new fraud patterns, create rules, evaluate performance, and improve your fraud metrics. We’re hands on before, during, and after implementation. You’ll have 24x7 support, an assigned account manager, and a dedicated Slack channel.

We’re at the forefront of fraud prevention

We are always one step ahead of the latest emerging fraud patterns, constantly shipping new features and optimizations to our platform. Our machine learning models are extremely effective, having been trained across billions of sessions to detect various fraud typologies.


Get started with an easy, low-lift integration

Our integration requires minimal engineering resources. We create a tailored plan for your integration and provide support at every step. Sardine will also configure your system and help you migrate rules from your existing provider.


An integration manager and sales engineer will be assigned to your project.

During the kickoff, we’ll walk through your integration plan and answer any questions.


We’ll support you throughout the technical integration process to help you get started as fast as possible.

Tech Review

We’ll review and test your integration. If everything looks good, we’ll issue your production credentials.

Go Live

To ensure a smooth go live, we’ll configure your account and train your team on the dashboard.

We’ll also monitor and validate your first sessions.


During the first month of live data, we’ll assist you in weekly rule and model performance evaluation.

After the first month, we’ll meet with you every other week to help you review fraud cases, create rules, and analyze performance.

Learn more about Sardine
Schedule demo