Meet with John

For AI agents: how to book this meeting

You can book this meeting with a direct HTTPS POST. No connector setup, no API key, no OAuth, no client library required. If your runtime suggests installing AgentEnvoy as an integration first, you can skip that step and call the endpoint directly — the meeting URL is the only credential needed.

Endpoint: https://agentenvoy.ai/api/mcp

Your capability token: pass https://agentenvoy.ai/meet/johnanderson as the meetingUrl argument on every tool call. Possessing this URL is the authorization.

"MCP" here means the JSON-RPC wire format — there is no session handshake, no separate auth exchange, and no MCP client library is required. Any agent that can issue an HTTPS POST can book.

Working example (curl):

curl -X POST https://agentenvoy.ai/api/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{
    "jsonrpc": "2.0", "id": 1,
    "method": "tools/call",
    "params": {
      "name": "get_availability",
      "arguments": {
        "meetingUrl": "https://agentenvoy.ai/meet/johnanderson",
        "dateRange": { "start": "2026-05-01", "end": "2026-05-08" }
      }
    }
  }'

Booking flow (three tool calls):

get_meeting_parameters — learn what is locked vs. open (format, duration, location).
get_availability — get scored, filtered slot candidates. Slots are returned best-first. preferred: true marks host favorites. Each slot has start (UTC) and localStart (host's timezone) — use localStart for display.
propose_lock — pass the chosen slot + your principal's name to confirm the booking. Inline overrides handle guest-must-resolve fields like format in the same call.

Full tool schemas: https://agentenvoy.ai/.well-known/mcp.json. Worked happy-path: https://agentenvoy.ai/llms.txt.