Skip to main content
A UI run in progress with live tiles and running counts
Every run has its own page under Test Runs. Watch it live while it runs, then open its report when it finishes: the pass rate, failures grouped by root cause, an AI-written summary, and a fix prompt for your coding agent.

​
Watching a Run

You don’t have to wait for a run to finish. The Run Report header at the top of the run page shows the pass rate and live counts (Running, Queued, Passed, Failed, Blocked, Total), refreshed every couple of seconds. The page shows results from this run, not each test’s latest status. The run page has two tabs showing the same run two ways:
The default for UI runs. A tile per test streams the agent driving your app (marked LIVE) and becomes a replayable recording when the test finishes. Queued tests join the gallery as soon as they start. See Agent Actions.
Agent Actions View with a running tile streaming live
Filter to Failed while the run is still going. A test appears under Failed as soon as it fails, so you can start triaging while the rest keep running.
Open any test to watch its steps land live: the step list fills in as the agent acts, and clicking a step shows an HTML snapshot of the page at that moment. See Step-by-Step Replay for how to read it.
Execution Steps list filling in live with a step snapshot in the right pane
A few things to expect while a run is live:
  • The live preview is a frame stream refreshed about once a second. The seekable video is produced when the test finishes.
  • Counts, steps, and snapshots refresh every 2 to 3 seconds. A blank snapshot usually means the page hadn’t settled yet; the next step shows the stable state.
  • Re-run all and Re-run failed stay disabled until the current run finishes.

​
Reading the Report

Click View Report on a run’s page to open that run’s report. It comes from an AI analysis pass that runs automatically when the run finishes; until it’s done, the page shows “Analyzing your results”.
View Report on a run's page, which opens that run's report
The Test Report page: stat cards, the pass/fail/blocked donut, and the Detailed Test Report button
The report covers outcomes and coverage, not timing; for when a run happened, use the Test Runs history.

​
The AI Fix Prompt

The Fix prompt for your coding agent card lists every concrete problem the run found and what to change. Click Copy prompt and paste it into Claude Code, Cursor, or any agent pointed at your repo. When there’s nothing actionable, the card is empty.
The Fix prompt for your coding agent card: a Copy Prompt button and the generated repo-oriented prompt
You’ll find the same kind of prompt in Issues to Fix on Insights and, for one failing test at a time, in the Step-by-Step Replay failure detail.

​
Downloading the PDF

The PDF is a fuller, print-ready version of the same run, with a cover, a table of contents, coverage tables, and per-failure detail including source code. Your browser produces it through its print dialog (nothing is stored on the server), so you can regenerate it any time.
1

Click Detailed Test Report

On the report page, click Detailed Test Report. A new tab opens and assembles the full report.
The Detailed Test Report button on the Test Report page
2

Let it finish generating

Wait for the Generating spinner to settle before printing.
3

Print to PDF with the right settings

In the print dialog, choose Save as PDF and use these settings so the report renders correctly:
Reports have no fixed expiry; they re-render from the run whenever you open them. Embedded screenshot links expire after about a week, so download the PDF if you need a permanent copy with images.

​
Where to Go Next

Insights

Issues to fix and quality trends across runs

Edit & Rerun

Re-run a test or the whole project after you’ve triaged