validate-ui

End-to-end UI validation for SSE streaming requires running browser automation against the live interface to verify heartbeats, reconnection behavior, and queue progression. You then compare UI state transitions against backend endpoints to confirm real-time correctness and identify streaming errors.

E2e testing for workflow execution loops and stale cancel behavior involves executing scripted browser snapshots across navigation flows to capture the UI state during parallel steps. It validates that the frontend correctly renders artifacts and handles cancellation or stale states as workflows execute.

A UX audit and codebase review for a dashboard and chat interface combines browser test execution with targeted review of critical UI modules. This produces a structured bug report evaluating correctness, accessibility, performance, and error handling across the tested navigation flows.

Yes, you need to launch both the backend and frontend locally to run end-to-end UI validation. The browser automation targets the local environment to test dashboard, chat, projects, and workflows, optionally seeding test data first to verify UI counts against backend endpoints.

Yes, you can use agent-browser scripted snapshots to verify EventSource and SSE infrastructure reliability. The validation process checks streaming UI state transitions, heartbeats, and reconnection behavior to ensure the real-time infrastructure handles errors and connectivity drops correctly.

The best way to test tool cards and error handling in a chat interface is through comprehensive browser automation combined with a focused codebase review. This approach validates UI rendering, streaming responses, and error states while generating a structured report of any UX gaps.