Miro MCP Server

What is Miro MCP Server?

An experimental Model Context Protocol (MCP) server that exposes Miro API endpoints as tools for AI agents, enabling programmatic manipulation of a single Miro board.

How to use Miro MCP Server?

Start by cloning the repository, installing dependencies (npm install), and creating a .env file with your Miro API token, board ID, and optional port. Then run npm run dev or build and start with npm run build and npm start. The server listens at http://localhost:8899/sse; AI agents connect via MCP to use the exposed tools. Docker is also supported via docker compose up -d.

Key features of Miro MCP Server

• Supports all major Miro board item types: sticky notes, shapes, text, images, cards, app cards, documents, embeds, and connectors
• Automatically generates tool definitions from the Miro OpenAPI specification
• Exposes board operations, item CRUD, and connector management as MCP tools
• Runs locally with Node.js or inside a Docker container
• Provides SSE transport for AI agent integration (default port 8899)

Use cases of Miro MCP Server

• AI agents programmatically adding sticky notes, shapes, and text to a Miro board
• Automating board layout updates, such as moving items or connecting them with connectors
• Extracting board data by retrieving items with filtering and pagination
• Enabling natural-language interactions with Miro through AI assistants

FAQ from Miro MCP Server

What is the Miro MCP Server?

An experimental MCP server that turns Miro API endpoints into tools for AI agents. It only operates on a single board specified by the user.

What prerequisites are needed to run this server?

You need Node.js, a Miro account with API access, a Miro API token, and the ID of the board you want to control.

What Miro board item types are supported?

Sticky notes, shapes, text, images, cards, app cards, documents, embeds, and connectors.

Can I run this server with Docker?

Yes. Use docker compose up -d (or docker-compose up -d depending on your Docker version) after configuring the environment.

Is this an official Miro product?

No. The README explicitly marks it as "experimental"—it is an unofficial integration using the public Miro API and OpenAPI specification.