Real-time simulation visualization and debug mode

Watch simulated conversations play out in real-time with parallel execution, versioned runs, and interactive debugging.

You can run multiple simulations in parallel, review versioned runs for comparison, and zoom into any conversation for detailed debugging. 

Enter debug mode to impersonate the simulated user and pinpoint exactly where issues occur in your conversation flow.

You can visually debug agent behavior, making it simple to collaborate with domain experts to identify edge cases in agent outputs and improve agent quality.

Get started: LangWatch Scenarios documentation →

Questions? Join our Discord community →