---
title: "Introduction to Agentic testing | Grafana Cloud documentation"
description: "Learn what Agentic testing is, when to use it, and how it compares to other Grafana Cloud testing options."
---

> For a curated documentation index, see [llms.txt](/llms.txt). For the complete documentation index, see [llms-full.txt](/llms-full.txt).

# Introduction to Agentic testing

> Caution
> 
> On October 26, 2026, the experimental **Agentic testing** app is removed from Testing &amp; synthetics. You won’t be able to view, edit, or re-run the tests it generated, and you can’t export those test definitions.
> 
> We’re rebuilding Agentic testing as a dedicated experience that works with Grafana Assistant. Recreate any journeys you still need before that date. Results from past runs stay in the Agentic Testing project until your retention window ends.

**Agentic testing** helps you validate critical user flows with functional browser tests in Grafana Cloud. You describe what should happen in plain language, and Grafana Cloud runs the journey in a browser and reports step-level pass/fail results with logs and a session recording.

**Agentic testing** sits alongside [Performance testing](/docs/grafana-cloud/testing/k6/) and [Synthetic Monitoring](/docs/grafana-cloud/testing/synthetic-monitoring/) under **Testing &amp; synthetics**. Use it when you want to confirm that important journeys work after a change without hand-writing browser test scripts.

> Note
> 
> **Agentic testing** is an [experimental feature](/docs/release-life-cycle/). Engineering and on-call support is not available. Functionality, UI, and availability may change.

## When to use Agentic testing

**Agentic testing** targets single-session functional checks: validating that critical user paths work after a change. Because you describe intent rather than implementation details, tests stay useful when button labels, layouts, or frontend frameworks change.

**Agentic testing** complements [scripted browser tests in Performance testing](/docs/grafana-cloud/testing/k6/get-started/run-your-first-browser-tests/), [k6 Script Authoring mode](/docs/grafana-cloud/testing/k6/author-run/k6-script-authoring-mode/), and [Synthetic Monitoring checks](/docs/grafana-cloud/testing/synthetic-monitoring/).

## Testing options

Use the approach that fits your goal:

Expand table

| Goal                                                       | Recommended approach                                                                                                                                                             |
|------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Validate a user journey without writing browser code       | **Agentic testing**                                                                                                                                                              |
| Write or generate k6 scripts for load or protocol testing  | [k6 Script Authoring mode](/docs/grafana-cloud/testing/k6/author-run/k6-script-authoring-mode/) or the [script editor](/docs/grafana-cloud/testing/k6/author-run/script-editor/) |
| Monitor endpoints continuously from global probe locations | [Synthetic Monitoring](/docs/grafana-cloud/testing/synthetic-monitoring/)                                                                                                        |
| Record or hand-write k6 browser scripts with full control  | [Run your first browser tests](/docs/grafana-cloud/testing/k6/get-started/run-your-first-browser-tests/)                                                                         |

## Test limits

The following limits apply when you create or run a test:

Expand table

| Limit                 | Value      |
|-----------------------|------------|
| Tests per stack       | No limit   |
| Steps per test        | 20         |
| Characters per step   | 300        |
| Maximum test duration | 15 minutes |

If a journey exceeds these limits, shorten step text, split the flow into multiple tests, or simplify checks before you save or run the test.

## Projects and access

**Agentic testing** stores test run data in a Performance testing project named **Agentic Testing**. Grafana Cloud creates this project the first time a [Grafana Cloud Admin](/docs/grafana-cloud/account-management/authentication-and-permissions/cloud-roles/) opens **Agentic testing**.

By default, only [Grafana Cloud Admins](/docs/grafana-cloud/account-management/authentication-and-permissions/cloud-roles/) can access the **Agentic Testing** project and use **Agentic testing**. To grant other users or teams access, a Grafana Cloud Admin updates permissions for the **Agentic Testing** project. Refer to [Manage user or team project access](../../k6/projects-and-users/configure-rbac/manage-roles-and-permissions/#manage-user-or-team-project-access) in Performance testing.

In **Testing &amp; synthetics** &gt; **Performance** &gt; **Projects**, the project appears as **Agentic Testing**.

## VUH usage and quotas

Each **Agentic testing** run is billed as a Performance testing test run and consumes virtual user hours (VUH) from your stack subscription. Refer to [Understand your Performance Testing invoice](/docs/grafana-cloud/platform/pricing-and-usage/performance-testing/) for how VUH is calculated.

To view consumption for the current month, go to **Testing &amp; synthetics** &gt; **Performance** &gt; **Projects** and check the **Quota** column for **Agentic Testing**. Refer to [Manage projects](../../k6/projects-and-users/projects/) and [Manage project limits and quotas](../../k6/projects-and-users/limits-and-quotas/) to learn how project quotas work and how Grafana Cloud Admins can change them.

## Experimental status and scope

While **Agentic testing** is experimental, availability may be limited to specific Grafana Cloud stacks or accounts, and UI labels, workflows, and supported journey types may change. **Agentic testing** targets functional browser journeys, not high-VU load tests or Synthetic Monitoring uptime checks.

## Next steps

- [Create browser tests from plain-language user journeys](../create-browser-tests-from-plain-language-user-journeys/)
