Adding API as a test category - #17

Can you explain how this works in practice? Does it crawl API endpoints in some way or does it need that array populated? Can you give some background on where this type of testing would be valuable as the APIs are reflected by page content and page functionality?

Can you explain how this works in practice? Does it crawl API endpoints in some way or does it need that array populated? Can you give some background on where this type of testing would be valuable as the APIs are reflected by page content and page functionality?

Usually we add API tests to automation frameworks because they fail faster than UI tests. We use them as a baseline for runs, if the whole API suite fails, there is a backend or network issue rather than a UI problem. In this framework specifically, API tests are also generated slightly faster than UI tests.

Since there's no Swagger file or external docs to generate tests from, the network traffic captured during crawling (_attach_network_listener) is used instead. The endpoints are filtered by the backend_url set in config (in the planner.py) and sent to the LLM to generate tests against. We also have a fallback if the AI didn't generate a test for a captured endpoint, a basic status check gets generated.

Good explanation I will review today