Monitor newly registered domain names around the world, automatically compare (including fuzzy comparison) suspicious domain names through machine learning technology, and perform risk assessments on them. Notify active phishing domains ( brands Party) to processing, continuous monitoring of inactive but suspicious domain names.

The search system is mainly used for preliminary screening and labeling of risk data. The underlying distributed architecture uses Kubernetes to implement automatic search of cross-platform host clusters based on Docker, and uses the cluster to monitor and allocate resources to resource sets. At the same time, for websites restricted by IP addresses, a springboard model based on IP addresses has been established to effectively prevent phishing scams and fraud techniques.

Integrate the search capabilities of mainstream search engines such as Google and Yahoo to ensure more comprehensive and accurate search results. Technically, MD5 hash value is used to ensure the accuracy and completeness of the comparison and achieve load balancing.

Different types of counterfeit assets have different degrees of dependence on the platform. For example, counterfeit APP assets are often attached to APP stores. In view of this feature, on the basis of other detection methods, more than 5,000 app platforms at home and abroad are scanned and detected, and newly added APP stores are continuously monitored, and suspicious assets are evaluated from multiple dimensions to quickly anomalies are found.

Model the image features of counterfeit assets and compare them with automatic screenshots of suspicious assets to score suspicious assets based on the risk assessment algorithm.