Agent API

    The agent API provides REST endpoints for controlling the agentic perception loop and querying observation memory. All endpoints are prefixed with /agent.

    POST /agent/start

    Start the agentic loop as a background task.

    Request body:

    {"goal": "monitor the scene"}

    Response:

    {"status": "started", "goal": "monitor the scene"}

    Returns 409 if an agent is already running.

    POST /agent/stop

    Stop the running agentic loop.

    {"status": "stopped", "ticks": 42}

    GET /agent/status

    Get current agent status.

    {
  "running": true,
  "tick_count": 42,
  "current_plan": ["Track person position", "Monitor hand proximity"],
  "goal": "monitor workspace safety"
}

    GET /agent/stream

    Server-Sent Events (SSE) endpoint that streams AgentTickEvent objects in real time while the agent is running. Each event contains observation data, reasoning, and the action taken.

    GET /agent/memory

    Get recent observations from the agent's memory store.

    ParameterTypeDefaultDescription
    limitint20Number of recent observations to return

    POST /agent/recall

    Query the observation memory store with structured filters.

    Request body: RecallQuery JSON — supports filtering by tags, time range, and significance threshold.

    GET /agent/demo/stream

    SSE stream of scripted demo data for frontend development without a running backend. Streams 10 pre-built ticks simulating a workspace safety monitoring scenario.

    ← Model AdaptersGovernance API →