Build with scheduling infrastructure

Tool Architecture

18 MCP Tools

Organized in 5 layers — discovery, temporal context, calendar operations, availability, and booking. Each layer builds on the one below. Your agent picks the tools it needs.

Reliability

Atomic Booking (2PC)

Two-Phase Commit prevents double-bookings. Lock the slot, verify no conflicts, write the event, release the lock. If any step fails, everything rolls back automatically.

Interoperability

Cross-Provider Merging

Google Calendar, Outlook, and CalDAV in one unified view. A single find_free_slots call merges availability across all connected calendars.

LangGraph

Use MultiServerMCPClient for tool discovery. Build reactive agents or multi-agent graphs with scheduling tools.

from langchain_mcp_adapters import (
    MultiServerMCPClient
)

async with MultiServerMCPClient(
  {"temporal-cortex": {
    "command": "npx",
    "args": [
      "@temporal-cortex/cortex-mcp"
    ]
  }}
) as client:
  tools = client.get_tools()
  agent = create_react_agent(
    llm, tools
  )

CrewAI

Add MCP tools to your crew's agents. The scheduling tools integrate as standard CrewAI tools.

from crewai import Agent, Task
from crewai_tools import MCPTool

scheduler = Agent(
  role="Scheduler",
  tools=[
    MCPTool(
      server="npx",
      args=[
        "@temporal-cortex/"
        "cortex-mcp"
      ]
    )
  ]
)

OpenAI Agents

Connect via the streamable HTTP transport. OpenAI Agents SDK discovers tools automatically.

from agents import Agent
from agents.mcp import (
    MCPServerStreamableHttp
)

server = MCPServerStreamableHttp(
  url="http://localhost:3001"
  "/mcp"
)
agent = Agent(
  name="Scheduler",
  mcp_servers=[server]
)