Your tests run in the vikingQA cloud on every release. Each run keeps its video and the full technical logs. When a test fails, nobody has to reproduce it to believe it, and the bug is caught before your customers see it. Run your tests on vikingQA, on your own infrastructure, or both.
Watch the run again, up to the moment it went wrong.
A batch is one execution of a selection of tests, on one environment, at one commit. Your pipeline, a schedule or a person starts it.
A recording of the browser for the full run. Jump to the moment a step failed.
The standard Playwright trace. Open it in the trace viewer to inspect the DOM, the actions and the timing at each point.
Each request as a HAR file, with status and timing. See the failed API call behind a failed step.
The errors and messages the page wrote during the run.
Each step of the test with its status and duration, in the words of the test case. Non-developers read the result without reading code.
The page at the moment of failure.
The output of the sandbox that ran the test, for the cases where the problem is not in the page.
Not every failed test is a bug. The agent reads the evidence first. If the test is out of date, the agent explores again and heals the test. If it is a real bug, ask the agent to create a ticket with the video and the trace, or to tell the team, in the tools you connected.
Nothing to do
A schedule runs a selection of tests on an environment at the times you set, in your time zone.
Nightly smoke on Staging: 40 passed, 2 failed
Two checkout tests failed at the payment step. Both passed yesterday. The network log shows an error from the payment service in both runs.
Some runs are automated. Some are verified by the agent. Some are run by a person with the browser extension. All three produce the same record: a verdict, the steps and the evidence.
Your release report shows the full picture: what was tested, how, and by whom.
Running your committed Playwright tests costs no AI credits. Credits pay for the agent's work: writing, exploring and healing tests. An AI run, where the agent verifies a test case that has no automated test yet, is agent work and uses credits.
Run in the vikingQA cloud, or in your own CI: the tests are plain Playwright in your repository. In the vikingQA cloud, each run has its own isolated sandbox. Our sandbox provider routes the work to an EU region, and it writes payloads larger than 2 MiB to object storage in the United States. Sandbox execution fully inside the EU is on our roadmap. Private runners inside your network are on our roadmap.
Yes. The trace is the standard Playwright trace file. Download it and open it in the Playwright trace viewer.
Retries follow your Playwright configuration. Each attempt is recorded with its own evidence, so you can see a test that passed only on the second attempt.
Yes. Invite them with the Customer role. They see the projects you share, and nothing else.
Not yet. Result and evidence push to Xray Test Executions and Zephyr Scale test cycles in Jira is on our roadmap.
Download the files. Evidence is kept for 90 days today. A longer or custom period per plan is on our roadmap.
Start free with your own application. No credit card.