// the find
authorjapps/zerocode
zerocode-tdd is a community-developed, free, open-source, outcome-driven automated testing for microservices APIs, Kafka(Data Streams), Databases and Load scenarios. all defined in simple JSON or YAML — with zero coding.
Zerocode is a declarative API and Kafka testing framework for Java teams that want to write test scenarios in JSON or YAML without writing test code. It sits between Postman collections and a full code-based test suite — you get repeatable, version-controlled tests with built-in assertion matchers, retry logic, and load testing. Real-world adoption by HSBC and UK Home Office suggests it handles production-grade complexity.
The assertion DSL is genuinely useful — STRICT vs LENIENT mode, regex matchers, date comparators, and array size checks without writing a single line of Java. Kafka support is first-class, not bolted on: you can produce, consume, and assert on message content with the same JSON syntax as HTTP tests. The load testing story (parallel runners, HighCharts HTML reports) is built in and doesn't require a separate tool. A published JSON Schema (Draft-07) for scenario files means IDE validation and AI-assisted test generation are practical today.
The DI layer is Guice, which means understanding why something fails often requires reading framework internals rather than your test file — debugging a misconfigured module is painful. Test data chaining between steps (using `${scenariolevel.step1.response.id}` style tokens) works but is fragile and opaque; complex multi-step workflows become hard to read fast. The README pushes a third-party CLI called Steply for 'language-independent mode' with no explanation of what that means or who maintains it — that's a trust red flag for a testing tool. No native support for GraphQL or gRPC, so if your stack uses either, you're on your own with the Java extension API.