Pangea API for Malicious URL Detection

API for malicious URL detection is a security measure that helps scan, identify and prevent malicious pages from spreading across your client/vendor sites. These malicious pages can be embedded in forums, comments, uploaded files and more. Detecting them early on can help you stop users from interacting with phishing/malware links in their accounts and prevent malicious user generated content that may lead to account compromise.

Several different algorithms exist to automatically scan a URL string and return a verdict on its maliciousness or benignity. Most use Natural Language Processing techniques such as bag-of-words. These methods assume that specific words or tokens are universally associated with malicious URLs and thus can be identified with high accuracy. These approaches can be limited by the size of the training set and fail to generalize to new URLs.

API for Malicious URL Detection: Prevent Phishing & Fraud

More sophisticated methods, such as supervised learning, take advantage of the fact that a URL type (malicious/benign) can be mapped to a point in space. A series of features are taken, such as the alphabetic distribution of vowels and consonants, and a straight line is then drawn between the points of malicious and benign URLs. Using a training set of data with labels for each feature, the algorithm can then accurately locate the point of an incoming URL and determine whether it is malicious or not.

The Pangea APIs offer a predictive model for URL classification that uses machine learning to look for non-benign indicators. This is a much faster approach than network-level fingerprinting, forensic analysis or visual inspection, and also offers a better chance of catching new threats as they emerge. It also avoids the drawbacks of blacklisting which can impose a large computational overhead and potentially limit a website/host to those with a known signature.