A multi-agent system where each person has their own AI assistant.
Agents coordinate care while respecting privacy boundaries.
Enter any username to start (creates account if new)
Loading...
Training & Feedback
π How Simulations Work
Simulations create synthetic users with realistic personas (anxious, forgetful, in denial,
etc.)
and run conversations to test the AI agent. Supports both regular scenarios (single care
circle)
and world simulations (multi-care circle dynamics with role inversions). Each simulation
automatically
generates feedback on response quality, escalation handling, and empathy. View results in
the "My Feedback" tab.
Feedback is captured via the dashboard form without blocking the CLI stream.
Network Visualization
Waiting for CLI participantsβ¦
π Simulation Complete
Deep System Analysis
Run advanced CLI analysis tools to optimize
prompts and identify failure patterns.
Simulation Progress Report
Generate and view interactive HTML reports
summarizing simulation performance.
π Training Metrics
Care Circle
My Circles
My Patients
Rate Last Interaction
How was the agent's response? This helps improve the system.