Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Testing and syntheticsbreadcrumb arrow Agentic testingbreadcrumb arrow Create browser tests
Grafana Cloud

Create browser tests from plain-language user journeys

Create a functional browser test by setting a start URL, adding plain-language steps, and saving the test in Grafana Cloud. For an overview of Agentic testing, refer to Introduction to Agentic testing.

Before you begin

To use Agentic testing, you need:

Contact your Grafana Cloud administrator if you do not see Agentic testing in the main menu.

Create a browser test

  1. Sign in to your Grafana Cloud stack.

  2. On the main menu, click Testing & synthetics > Agentic testing.

  3. Click + Create a new test.

  4. In Start URL, enter the full URL where the browser opens first.

    For example, enter https://quickpizza.grafana.com/admin. QuickPizza is Grafana’s demo site, so you can run this example without setting up your own application.

  5. In User Journey, add one action or check per step in plain language. Steps run in order after the start URL opens.

    For example:

    text 
    1. Fill username with admin
2. Fill password with admin
3. Click Sign in
4. Verify the page shows Latest pizza recommendations

  6. In Test name, enter a name that describes the journey, for example QuickPizza admin login.

  7. Click Save, or click Save and run to save the test and start a run immediately.

To store logins, API keys, or other sensitive values, use Grafana Secrets management and reference them as $SECRET_NAME in your journey steps.

Run a browser test

If you clicked Save and run when you created the test, Grafana Cloud starts the run immediately. Otherwise:

  1. Open the test you created.
  2. Click Run.

Agentic testing runs one functional browser session on Grafana Cloud. Each step executes in a real browser, not as a protocol-only or uptime check.

While the test runs, you can follow live progress in the run view, including the current step and browser logs.

Review test results

When the run finishes, review the results to confirm whether the journey passed or failed.

Each run includes:

  • Step-level pass/fail results for every step in the user journey
  • Logs that show what happened during the run
  • A session recording you can replay to see the browser interaction

Use these artifacts to debug failures, share outcomes with your team, and iterate on the journey description without rewriting a full browser test script.

For additional context on browser result views in Performance testing, refer to Inspect browser test results.