Skip to main content
These pages describe how to write, configure, and run Sedum tests.
  • Plain-English browser tests: the basics of e2e tests written in plain English, and where Sedum fits.
  • Probabilistic testing: how scores become verdicts, flags, and exit codes, and how to tune thresholds.
  • TypeScript tests: *.test.ts files with test(), ai steps, values and secrets, groups, extract, and Playwright code between steps.
  • YAML format: the *.test.yaml and *.module.yaml format, data, placeholders, modules, and hooks.
  • Project configuration: sedum.config.yaml, named environments, .env loading, test discovery, and precedence.
  • CLI commands: sedum init, doctor, run, validate, list, and browsers install, with options, output, and exit codes.
  • Running in CI: JUnit output, job summaries, and GitHub Actions, GitLab, and Jenkins examples.
  • Run results: the RunResult schema, live progress, reporters, and evidence privacy.
  • Assertion engine: verify and measure results, verdict policy, observation errors, and page limits.
  • Step classification: supported operations, offline validation, and the classification cache.
  • Local locator cache: development and CI defaults, parallel runs, privacy, and clearing the cache.
  • TypeSafe provider: provider setup, data sent to the provider, and the opt-in live check.
  • Clef provider: Cloudflare credentials, models, privacy, pricing, and the text-only Phase 1 scope.