Skip to main content

Integration Test Generation

RoostGPT generates integration tests that validate end-to-end workflows by combining your API specification with Gherkin feature files. Each Gherkin scenario becomes a complete integration test with supporting test data.

How It Works​

  1. Inputs required — An API specification (OpenAPI/Swagger) and a Gherkin .feature file (typically produced by the Functional Test phase)
  2. Scenario parsing — RoostGPT extracts each Scenario or Scenario Outline from the feature file
  3. Test generation — For each scenario, a complete integration test is generated using the API spec as the service contract
  4. Data generation — CSV test data files are automatically created for data-driven testing
  5. Output — A structured test suite with one file per scenario

Inputs​

InputSource
API SpecificationGenerated by Functional Test phase, or your own Swagger/OpenAPI file
Gherkin Feature FileGenerated by Functional Test phase, or your own .feature file

Supported Frameworks​

  • Karate
  • Rest Assured
  • Postman

Using the Web UI​

  1. Navigate to Create Test → Integration Test
  2. Select Test Framework (e.g., Karate)
  3. Configure your AI Provider and Git Repository
  4. Set Business Requirement Source to your Jira ticket or upload a Gherkin file directly
  5. In Gherkin Setup, choose the source:
    • File Upload: Upload your .feature file
    • URL: Link to Gherkin spec
    • Git Path: Repository path to the .feature file
  6. Save and execute

Using the VS Code Plugin​

  1. Right-click on your Swagger/OpenAPI spec file
  2. Select "Generate Test using RoostGPT" → "Integration Test using API spec and gherkin feature"
  3. Select your output framework
  4. Browse to and select your Gherkin .feature file
  5. Click Generate

RoostGPT parses each scenario from the feature file and generates corresponding test files.

Output Structure​

integration-tests/
├── CreateCard_scenario.feature         ← per-scenario test file
├── CreateCard_scenario_data.csv        ← test data for data-driven testing
├── RetrieveCard_scenario.feature
├── RetrieveCard_scenario_data.csv
└── ...

AI Provider Recommendations​

  • Claude Sonnet 4 — best for understanding complex Gherkin scenarios
  • GPT-4o — reliable for most integration test workloads
  • Gemini 2.5 Pro — handles large feature files well (1M context)

Next Steps​