
Cucumber
Automation Frameworks
Open source
Platforms
Web
API
iOS
Android
Languages
Java
JavaScript
Ruby
<What you get/>
Readable scenarios your product and business stakeholders can review before the build
A thin step layer that delegates to page objects, so automation changes don’t rewrite the specification
A living specification that stays accurate because it executes
Guidance on which behaviour belongs in Gherkin and which belongs in faster lower-level tests
What it is
Cucumber executes scenarios written in Gherkin by mapping each plain-language step to automation code. The value is not the tooling but the shared artefact: analysts, developers and testers argue about behaviour in one document, and that document is what runs. It works where teams genuinely collaborate on scenarios, and adds overhead where they do not.
How PerfectQA uses Cucumber
We use Cucumber when the people deciding what ‘correct’ means aren’t the people writing the code. Scenarios in Gherkin give a product owner or business analyst something they can read and dispute before it is built, which catches specification defects far earlier than a test run does. The discipline that makes it work is keeping step definitions thin — a step maps to one intention and delegates immediately to a page object or service layer, so the readable layer and the automation layer can change independently. Where we see Cucumber fail on inherited projects, it is almost always because steps became a scripting language and scenarios turned into click-by-click instructions nobody reads. We keep scenario counts deliberately low and reserve them for behaviour worth agreeing on in writing, running the bulk of regression coverage underneath at a lower level.
Category
Automation Frameworks
In our stack
6 years
Projects
22 delivered
Frequent questions
Does BDD slow us down?
Who writes the scenarios?
Can we use Cucumber with our existing Selenium suite?
What if nobody reads the scenarios?
Every stack is different
Tell us yours, and we’ll show you where this fits





