Use cases
Test cases from API specs, for API teams
Provo drafts API test cases from your spec: status and error codes, request validation, authentication, idempotency, retries, timeouts and ordering, only as the spec states them. Each case quotes the requirement it tests. Anything the spec leaves open, such as a backoff interval or what a redirect means, is listed as a question instead of being guessed.
What you get from an API spec
- Negative cases for every rule the spec states: wrong scheme, missing key, replaying a delivered event.
- Edge cases at the limits the spec gives: exactly 5 retries, an answer after 10 seconds.
- Regression cases for behavior the spec says must keep working.
- Open questions for vague or missing details, quoted from the spec.
Formats your tools can read
Export the suite as Gherkin .feature files for BDD tools, CSV for spreadsheets and test-management imports, JSON for your own scripts, or Markdown for a pull request or wiki. Each case carries its requirement key and quote.
What Provo doesn't do for APIs
Provo doesn't call your API, read OpenAPI files as a schema, or import automated test results. It plans and tracks the cases; your test framework runs the automated ones.
Questions
Can I paste an OpenAPI file?
You can paste or upload it as text, and Provo will split what reads as requirements, but it isn't parsed as a schema. Prose specs and tickets give better traces.
Try it on an API sample?
Yes: the no-account sample includes a webhook delivery spec for a fictional API, with its precomputed suite.