Verifying the correctness of Web APIs is an expensive, tedious task.
For example, writing test cases to send HTTP requests and verify the responses returned from the APIs requires expertise and time.
Automatically generating test cases to find functional faults, security vulnerabilities and cover requirements would hence be highly beneficial.
You can throw an OpenAPI schema to your LLM of choice, and get some test cases out of it.
Results will be shallow, though.
In the last few years, advances in AI techniques have led to few specialized open-source tools that can be used for this task, like EvoMaster, CATS, Restler and Schemathesis.
But most testers and developers in industry are still not aware of these existing solutions.
This talk will summarize the current state-of-the-art on this exciting testing automation problem, discussing their strengths and limitations.
We will briefly as well discuss our direct experience in introducing these techniques in large enterprises such as Meituan and Volkswagen.