Free tools

Playwright JSON Report Viewer

Open the file Playwright's JSON reporter writes: failed and flaky tests per project, every retry, and error messages with the terminal colour codes removed.

Runs in your browserNothing is uploadedNo sign-up

Paste the file, or choose or drop one or more files. Reads JUnit XML, TRX (Visual Studio / dotnet test), NUnit 3 XML, Playwright JSON, CTRF JSON.

Or drop files on the box. Nothing is uploaded.

How to get the JSON report

Playwright's JSON reporter prints to standard output unless you name a file:

npx playwright test --reporter=json > results.json
PLAYWRIGHT_JSON_OUTPUT_NAME=results.json npx playwright test --reporter=json

Or set it in playwright.config.ts, next to the HTML reporter if you use one:

reporter: [['json', { outputFile: 'results.json' }]],

The JSON file is the one to keep when a script or a dashboard needs the results. The HTML report is for people, and the blob report exists to merge the reports of sharded runs with npx playwright merge-reports.

What the viewer reads

The JSON nests suites (files, then describe blocks) down to specs, and each spec holds one test per project. The viewer flattens that:

  • One row per spec and project. With chromium and firefox projects, "adds an item" appears twice, marked with the project, because one can pass while the other fails.
  • Status comes from the test's outcome. expected is a pass, unexpected a failure, skipped a skip, and flaky a test that passed on a retry, shown as flaky with its retry count. When the last attempt timed out or was interrupted, the failure counts as an error.
  • Time is the sum of every attempt, so a flaky test shows what it really cost the run.
  • Errors lose the terminal colour codes that Playwright keeps in the JSON, so Timeout 5000ms exceeded reads as text.

Finding flaky tests across runs

A test that is flaky within one run shows up here as flaky. A test that passes in one CI run and fails in the next, with no retry in between, only shows when you compare runs: drop the JSON from several runs into the flaky test finder.

To split a slow suite, Playwright's --shard=1/4 splits by test file, or by individual test when fullyParallel is on. The test shard balancer reads this JSON and shows how evenly your files can be divided.

Sources

Testland for teams

In development

Automated test reports, manual test runs and regression tracking in one workspace. A separate product, built on the same skills.

One email when it opens, nothing else. Privacy policy