Runs & artifacts

A run executes a feature's generated tests in an isolated cloud sandbox and leaves behind a results digest and downloadable artifacts.

A run is one execution of a feature's generated Playwright tests, in an isolated cloud sandbox, against the project's URL (or a siteUrl override).

Run lifecycle

Dispatching a run returns a runId immediately; the sandbox spins up and executes in the background.

Status	Meaning
running	Queued or executing in the sandbox.
passed	Finished, every test passed.
failed	Finished with failures — the digest carries per-test errors and a trimmed runner log.

Typical wall-clock time is ~30 seconds for a small feature. Poll with get_run or block with wait_for_run / testvibe run … --wait.

The results digest

get_run returns a digest designed to be read by humans and AI tools:

{
  "runId": 214,
  "status": "passed",
  "featureName": "Login Functionality",
  "config": "Default",
  "summary": { "total": 3, "passed": 3, "failed": 0, "other": 0 },
  "tests": [
    { "name": "Valid user signs in", "status": "passed", "durationMs": 4210, "error": null }
  ],
  "artifacts": [
    { "kind": "trace", "path": "trace/login.zip", "size": 482133 }
  ]
}

On failure, each failing test carries a trimmed error, and the digest includes a log excerpt from the runner.

Artifacts

Runs upload Playwright traces, screenshots, videos, and reports to artifact storage. get_run_artifacts exchanges the manifest for short-lived signed download links.

Artifact URLs expire quickly by design — fetch them right before downloading, and don't persist them.

Configurations

A run executes under a named Playwright configuration (Default unless you pass config) — that's where browser, viewport, and timeout choices live.