Ensuring the reliability of APIs is critical in distributed systems and DevOps workflows. Chaos engineering has emerged as an effective approach for identifying vulnerabilities, yet its application in API testing remains limited due to the lack of automation and efficiency in generating diverse test cases. This paper introduces ChaosAPI, a novel tool leveraging genetic algorithms to automate chaos engineering in API testing. Unlike traditional methods, ChaosAPI uses OpenAPI Specifications to systematically generate, mutate, and optimize test cases, ensuring comprehensive test coverage and the detection of unhandled edge cases. A comparative study demonstrates ChaosAPI’s effectiveness against state-of-the-art methods, including NLP-based testing techniques, Llama 3 8B, and JetBrains AI Assistant. Despite requiring less computational power and having no dependency on source code access, ChaosAPI achieves test generation performance comparable to advanced AI-driven tools.