How to let your AI agent manage your calendar with KalendMe MCP
KalendMe
AI agents already draft emails, summarize documents, and answer questions. But ask your agent "Am I free on Thursday afternoon?" or "Book a 30-minute call with this client" and it hits a wall: it can't access your calendar.
KalendMe now ships an open-source MCP server that gives any AI agent direct access to your scheduling.
What is MCP?
The Model Context Protocol (MCP) is an open standard for connecting AI applications to external tools and data sources. Instead of building a custom integration for every AI assistant, you expose one MCP server and any compatible client can use it.
KalendMe's MCP server is published as an npm package (@kalendme/mcp-server) and available on GitHub.
What can your AI agent do with it?
Once connected, your agent gets access to 18 scheduling tools:
- Check availability: "When am I free next week?"
- Book sessions: "Schedule a 30-minute call with maria@example.com on Monday at 2pm."
- List and manage booking links: "Create a new 45-minute consultation link."
- View and cancel events: "Cancel my 3pm meeting tomorrow and notify the guest."
- Block time: "Block Friday afternoon for deep work."
- Manage webhooks: "Set up a notification when someone books a session."
Your agent handles the conversation. KalendMe handles the scheduling logic: conflict detection, timezone conversion, calendar sync, and email notifications.
Getting started in 3 steps
1. Get your API key
Log in to KalendMe, go to Settings → API Keys, and generate a new key. You need a Premium plan for API access.
2. Install the MCP server
If you are using Claude Desktop, add this to your claude_desktop_config.json:
{
"mcpServers": {
"kalendme": {
"command": "npx",
"args": ["-y", "@kalendme/mcp-server"],
"env": {
"KALENDME_API_KEY": "your-api-key-here"
}
}
}
}
For other MCP-compatible clients (like NanoClaw or OpenClaw), the setup is similar. Point the client at the @kalendme/mcp-server package and provide your API key as an environment variable.
3. Start talking to your calendar
That is it. Your AI agent can now answer scheduling questions and take actions on your behalf:
You: "Am I available for a 1-hour meeting next Wednesday afternoon?"
Agent: "You have two open slots on Wednesday: 2:00–3:00 PM and 3:30–4:30 PM (America/Sao_Paulo). Would you like me to send a booking link?"
You: "Book the 2pm slot with ana@company.com, title it 'Project kickoff'."
Agent: "Done. I booked 'Project kickoff' on Wednesday at 2:00 PM. Ana will receive a confirmation email with the calendar invite."
Real-world use case: a personal scheduling assistant
Imagine you run a consulting practice. You have a NanoClaw or OpenClaw agent connected to your WhatsApp. A potential client sends you a message:
Client: "Hi, I'd like to schedule a consultation. Are you free this week?"
Your agent checks your KalendMe availability, offers the open slots, and books the meeting. The client gets a confirmation email, the event appears on your Google Calendar, and your availability updates automatically. No manual steps.
Why MCP instead of a REST API?
KalendMe already has a full REST API. So why MCP?
The difference is the audience. The REST API is for developers building applications. The MCP server is for AI agents. It provides structured tool definitions, input validation, and natural-language-friendly descriptions that help the agent understand what each tool does and when to use it.
You don't need to write any code. You just connect your agent and start talking.
Open source and extensible
The KalendMe MCP server is fully open source. You can inspect the code, contribute improvements, or fork it to add custom tools specific to your workflow.
Install it from npm:
npm install @kalendme/mcp-server
Or run it directly:
npx @kalendme/mcp-server
Start scheduling with AI today
If you already use KalendMe, you are one configuration step away from giving your AI agent full calendar access. If you are new to KalendMe, create a free account and upgrade to Premium to unlock API and MCP access.
Your calendar should understand natural language, not just clicks.
KalendMe