Miro MCP Server overview

Who can do it: All Miro users
Which plans: All plans
Which platforms: Browser, Desktop

Miro's Model Context Protocol (MCP) Server is a gateway that allows AI tools to connect with Miro boards, enabling seamless interaction between your visual workspace and external AI systems. This powerful integration lets AI assistants read from and write to your Miro boards, creating a bridge between visual collaboration and AI capabilities.

What is Miro MCP Server?

Miro's MCP Server is a secure gateway that gives AI tools secure access to Miro boards within your organization. It enables AI assistants to:

• Read and understand content from your Miro boards, including frames, documents, prototypes, tables, and diagrams.
• Create and update boards, diagrams, documents, tables, and code widgets.
• Generate diagrams based on code, text descriptions, or a pull request.
• Transform visual information into code and other formats.
• Read, reply to, and resolve comment threads on a board.
• Work with images, including uploading and retrieving them from a board.

TIP: Think of MCP as a universal translator that helps different AI systems communicate with Miro.

MCP Server + Clients

MCP provides the underlying infrastructure, while MCP Clients are specific integrations with AI tools:

• MCP Server: The secure gateway that enables communication between Miro and external AI tools.
• MCP Clients: Specific implementations that connect particular AI tools (like Claude, Cursor, VSCode, or Lovable) to Miro.
• OAuth 2.1 Security: Enterprise-grade authentication ensures secure access to your content.

NOTE: Enterprise accounts require admin approval before users can install and use Miro's MCP Server.

How it works

1. Connection:
   1. In Enterprise plan, an admin enables Miro MCP for their organization or team.
   2. In all other plans, MCP access is enabled but needs to be authorized via an MCP Client (i.e. Cursor, Claude Code etc).
2. Authorization: Users connect their preferred AI tools to Miro through OAuth.
3. Interaction: AI tools can then read from and write to boards the user has specified.
4. Creation: Users can request AI tools to generate diagrams, documents, tables, code widgets, and more based on board context.

NOTE: For more technical details and examples, visit our developer documentation.

MCP tools

Miro's MCP Server exposes a set of tools your AI agent can call, grouped into a few categories:

• Boards: Create new boards, search boards you have access to, and read items on a board with pagination and filtering.
• Diagrams: Create diagrams from a DSL description and read the DSL layout format.
• Documents: Create, read, and update structured document items on a board.
• Tables: Create tables with defined columns, and add or update rows.
• Code widgets: Create, read, update, and delete code widgets on a board.
• Comments: Read comment threads, reply to them, and mark them resolved.
• Images: Upload images, and retrieve image data or a download URL for an existing image item.
• Layout: Read and update board items in bulk using a DSL representation, and get the layout format specification.
• Prototypes: Read existing prototype screens, and create a prototype from one or more HTML screens.

NOTE: For the complete, up-to-date list of tools and prompts, including which ones use Miro AI credits, see the Miro MCP tools and Miro MCP prompts.

Usage limits

Each Miro plan includes a daily allowance of MCP tool calls, which resets at 00:00 UTC:

If you reach your daily allowance, further MCP tool calls are blocked until the limit resets, and your AI tool receives an error message. Enterprise admins can request a higher allowance under Admin settings > Apps and integrations > MCP.

NOTE: For full details, see the Miro MCP usage and daily limits documentation.

Explore the Miro MCP world

• Check our developer documentation on MCP.
• Check out this youtube playlist with fun tutorials on what you can do with Miro MCP.

More information: See What are MCP Clients and How to enable Miro's MCP Server (user guide).