1.21.0 — Reusable test data & the Agent Home cockpit

This release introduces a tenant-wide catalog for reusable test data, rebuilds Agent Home into a genuine cockpit for the autonomous AI agent, and makes scripts testable in isolation.

Reusable test data catalog. Maintain test data once — as a dynamic table or CSV — in its own Settings section and bind it to test-case variants across many tests, instead of copying values from test to test. See Test Data.

Agent Home: plan health and the needs-you queue. The redesigned Agent Home shows a traffic light per execution plan under the AI agent's watch, a needs-you queue that explains in the agent's own words why an item needs you — with Keep, Revert, Accept and Discuss actions — and a "What I did" timeline with an inline health chart for the last days. See Agent Home.

Try a script live. A test this script dialog runs a script on demand against your application — with the login derived from the selected test user — and shows the captured logs when it fails, so you no longer need a scaffolding test case to develop a script.

New welcome landing page. A new home page greets you with quick actions and, for applications monitored by the AI agent, embeds the agent dashboard so suite health is the first thing you see.

Localized execution screenshots. Screenshot descriptions in execution results are now translatable and use clearer, action-based step labels. See Execution.

Polish and fixes. German texts use consistent wording throughout (an execution plan is now called Ausführungsplan everywhere), the back button on detail screens follows your browser history instead of a fixed route, and dashboards adapt their tile layout to the available width.