What it does
The simulator enforces the signal constraints; its public demo uses an adaptive fallback without Jev credentials, so the moving map alone does not prove live inference.
Playable demos & bots
Steer a Chicago traffic simulation with Jev choosing citywide priorities and bounded corridor policies.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
The simulator enforces the signal constraints; its public demo uses an adaptive fallback without Jev credentials, so the moving map alone does not prove live inference.
Your developer can study this project to learn how to guide an outside AI safely. The code gives the AI a strict list of traffic priorities to choose from. The AI picks a general goal, like easing citywide pressure. Then, your local program safely controls the actual street lights.
The public example runs without an access key that connects to the AI service. It uses a backup routine instead. This routine keeps the map moving when the real AI is not connected. Your developer can borrow this method to keep your app working during service outages.