Systematic Literature Review: RESTful API Testing Using Machine Learning Techniques

Abstract

Representational State Transfer (REST) APIs are central to modern web and microservice systems, making their reliable testing increasingly important. Traditional automated approaches—black-box, white-box, and model-based have improved coverage but still struggle with incomplete specifications, invalid inputs, and complex workflows. Recent advances in machine learning and large language models address these issues by extracting constraints, generating realistic inputs, and guiding adaptive exploration. This review outlines how machine learning and large language models are transforming REST API testing, summarizes key techniques and empirical results, and highlights open challenges and future directions toward more intelligent, automated, and scalable RESTful API testing solutions.