Task Visualizations
As agents work on tasks, they automatically capture screenshots and screen states from applications they interact with. These visualizations show exactly what the agent saw during execution—giving you visibility into how the agent tested and verified its changes.
What Gets Captured
Agents capture visualizations whenever they interact with running applications:
- 5250 terminal screens: Character-based displays from IBM i systems and legacy applications
- Rich Display screens: Modern web interfaces built with Profound UI or similar frameworks
- Web page screenshots: Browser-rendered pages the agent navigated to
Captures happen automatically during task execution. Agents know when to capture screens based on the testing context and instructions configured in the environment.
Viewing Visualizations
After a task completes, you can view the captured screens:
- Open the task detail page
- Look for captured screens in the Output Summary section
Each visualization shows:
- The rendered screen as the agent saw it
- When it was captured during execution
- Which session or test produced it
For 5250 screens, visualizations render in a terminal-style format with proper formatting and colors. For web interfaces, you see the page as it appeared in the browser.
Why Visualizations Matter
Visualizations help you:
- Verify agent work: See exactly what the agent tested, not just what it reported
- Debug issues: When something goes wrong, screenshots show the application state
- Document changes: Captured screens serve as visual evidence of testing
- Understand legacy systems: 5250 screens are rendered in a readable format even if you're unfamiliar with terminal interfaces
Configuration
Visualizations require environment-level configuration to render certain screen types (especially 5250 screens). If visualizations aren't appearing for your environment, ask your administrator to check the environment settings.
See Environments in the Administration section for configuration details.