/** * OpenKernel — RoboPark Control dashboard (served by a mesh node). * * A single self-contained page (no external assets — CSP-clean, ships in dist as * a compiled string). It polls GET /fed/status, renders the fleet + this node's * hardware + a live activity feed, and runs allow-listed actions via * POST /fed/command and POST /fed/apply. The browser authenticates every request * with the ?token= it was loaded with (browsers can't set Bearer on navigation). * * Three tabs: * Fleet — node cards + this-node hardware + live activity feed * Park — isometric live map of the mesh (nodes placed on a floor, agent * movement + tool-use pulses streamed from /fed/stream; click a robot * for its cam / mic / speaker / hardware drawer) * Sessions & Telemetry — RoboPark scheduler stats + production_mode + controls * * The Park cam panel relays RoboVisionAI_PI's existing MJPEG feed through the * authenticated mesh. LiveKit remains the voice transport but is deliberately * not used for dashboard video, avoiding camera ownership conflicts on robots. * * Kept free of backticks and ${...} so the whole page fits inside this TS * template literal without escaping; the client JS uses string concatenation. */ export declare const DASHBOARD_HTML = "\n\n
\n\n\n\nLive readiness, workload, release compliance, hardware pressure, and operator actions across every RoboPark device.
| Robot | Session | Started | Duration | End reason | Actions | \n
|---|
Configure a call or copy a reusable, credential-free ElevenLabs join link directly from a character card.
| Session | Robot | Started | Ended | Duration | End reason | Turns | Pipeline | Latency | Actions | \n
|---|
| Time | Category | Action | Target | Details | Operator | \n
|---|
| Container | State | Image | Actions |
|---|
Set up or repair a complete robot in about one minute. Choose three things, copy one command, and paste it on the robot.
The command installs RoboPark and creates the unified mesh, camera, audio, motion, motor, and RoboVoice hardware runtime. Direct ElevenLabs agents use the separate commands below because the two engines must not own the same media devices at once.
Build reusable personas and production voice pipelines in one place. Changes are applied centrally and become available immediately in Voice Studio, Fleet, and every Park robot panel.
Speech recognition, language model, synthesis, turn-taking, and wake behavior.
Persona, prompt, robot avatar, park placement, motors, and default voice stack.