Reports and analytics
Inspect execution evidence and compare results and trends under matching conditions
A report describes what a run checked and what happened. Analytics summarizes accumulated execution records to help find recurring failures and changes.
Read a run
Open Verification → Run history and first check its tests, environment, persona, browser, and execution time. Review skipped checks and interrupted runs alongside passes and failures.
A pass rate describes executed checks. It is not a product-wide quality score covering untested features or missing assertions.
Inspect failure evidence
A selected test has Content, Run history, and Artifacts views. Expand code and documents within Content. Register findings as improvements to retain evidence and subsequent retest decisions.
Use report details and debugging artifacts to inspect the failure. Available evidence depends on test configuration and successful collection.
| Evidence | What to inspect |
|---|---|
| Errors and execution logs | Failed step and expected outcome |
| Screenshots | Screen state at capture time |
| Console and network records | Browser errors, failed requests, and response status |
| Playwright traces | Recorded execution steps and page states |
| Visual comparisons | Differences between baseline and current capture |
| Accessibility and performance results | Findings and metrics collected by those checks |
Treat AI investigations as hypotheses. After a fix, rerun under the failing conditions to verify it. Before approving an intentional visual change as a new baseline, check that it does not include an unrelated regression.
Compare runs
In Changes → Test results, select two different runs of the same test and open Compare results. See Changes for comparison conditions.
Match the test scope and execution conditions. Separate new failures, resolved failures, and newly added checks. If the environment or test code changed, product changes alone may not explain the difference.
The report's rerun action opens run settings with that test selected. It uses current code and conditions; it does not restore a historical execution.
Review accumulated results
Choose Verification → Check type → Run statistics for success-rate trends, execution times, error categories, failure patterns, browser comparisons, and unstable tests. Period comparisons need similar test composition and execution volume to be meaningful.
Switch projects to inspect each service's results. The website's multiple-project illustration is not evidence of automatic testing or a live aggregate of every project. A project needs execution records before it has meaningful analytics.
Share results
Individual report exports support CSV and Markdown. The report list provides batch CSV and ZIP exports. Review account details, URLs, screenshots, and logs for sensitive content before sharing.
Use MCP collaboration to let external AI read the latest results, or Scheduled checks to collect results at intervals.
Overview dashboard
Quality overview loads compact metrics for the entire retained history and displays 7, 30, or 90 days. Detailed logs and artifacts are excluded from this response. Runs, pass rate, and failed checks are compared with the immediately preceding period of equal calendar length. Today includes only elapsed time; deleted reports are excluded.
- Execution trend shows daily passed, failed, and skipped checks, with a range brush.
- Test results shows the outcome distribution. Pass rate is passed / (passed + failed), excluding skipped checks.
- Recurring failures ranks the five tests with the most failed checks, not the highest failure rate.
- Improvement status uses all current improvement records, independently of the report period.
Widget settings controls visibility, reading order, and width. Use the arrow buttons to reorder charts and the width button to choose a standard or wide desktop panel. Small screens use one column. Settings are saved per project on this device and can be reset to the default layout. Expand the daily aggregation table for numeric data. Empty data and request failures are distinguished; chart details open the corresponding execution evidence or improvement workspace.
Evidence and coverage
Click run count, pass rate, or failed checks to inspect the executions used in the current period. Outcome buttons, test names below recurring failures, and dates in the daily table narrow the evidence list. Search stays within that selection. Open record navigates to the exact test and execution timestamp.
Feature and flow coverage uses each linked test's latest retained execution, independently of the period filter. It distinguishes unlinked tests, deleted test references, missing execution evidence, recent failures, skipped checks, and recent passes. Every linked test needs evidence before an item can be classified as passed. This describes linked tests, not an overall product quality score. Status buttons filter the list; names open product structure details and evidence buttons open the corresponding run.
Execution configuration filters
Environment, browser configuration, and test framework filters apply together to charts, previous-period comparisons, and execution evidence. Feature and flow coverage continues to use the complete retained history independently.
New executions retain explicitly selected environment and browser names, never environment variable values. Older records or runs without an explicit selection appear as No configuration recorded. Reports combining several browsers remain grouped by that combination because their results are aggregated. If a report cannot be read, its excluded count is shown.