Connecting AI coding assistants with Miro using the Miro MCP Server

Connect your AI coding assistant with Miro using the Miro MCP Server to bring board context into your development workflows. This guide covers two clients — GitHub Copilot CLI and Claude Code — but the same server works with any MCP-capable tool.

Once connected, your assistant can read Miro boards, summarize diagrams and notes, and help generate code or documentation from the ideas captured on the board.

What is the Miro MCP Server?

The Miro MCP Server is Miro’s implementation of the Model Context Protocol (MCP) that connects Miro’s collaborative design platform with AI assistants like GitHub Copilot and Claude Code.

It enables:

• Reading board context — AI can analyze your Miro boards, designs, and PRDs
• Generating diagrams — Create flowcharts, UML diagrams, ERD schemas based on code or text
• Generating code — Transform PRDs, diagrams, and prototypes into working code
• Manipulating board items — Create sticky notes, shapes, text, images, cards, documents, and more
• Enterprise security — OAuth 2.1 authentication with secure access control

See the Miro MCP Introduction for more details.

Prerequisites

• A Miro account with MCP enabled for your organization (if on Enterprise, your admin must enable the MCP Server first)
• An MCP-capable client installed — for example Copilot CLI or Claude Code
• Access to the Miro board you want your assistant to use

Miro MCP Server OAuth 2.1 Authentication

Miro’s MCP Server uses OAuth 2.1 with dynamic client registration, which means:

• No API tokens to manage — authentication is token-based and expires automatically
• User-level consent — you grant permission through a browser-based flow
• Enterprise-grade security — compliant with enterprise security standards
• Dynamic registration — the MCP server automatically registers your client

Configure Copilot CLI with Miro MCP Server

Step 1: Add the Miro MCP Server configuration

Add the Miro MCP server to your Copilot CLI configuration file at ~/.copilot/mcp-config.json:

{
  "mcpServers": {
    "miro-mcp": {
      "type": "http",
      "url": "https://mcp.miro.com/",
      "tools": ["*"]
    }
  }
}

Step 2: Start the Copilot CLI and grant access via OAuth flow

Start the Copilot CLI and follow these steps during the OAuth flow:

1. Review the OAuth prompt — A browser window opens to Miro’s authorization page
2. Select your team — Choose which Miro team to grant access to - This determines which boards and projects Copilot can access
3. Confirm permissions — Review and approve the requested permissions
4. Continue — You’ll see a confirmation and be redirected back to your terminal

Once authorization is complete, Copilot CLI caches the access token and subsequent commands use it automatically.

Step 3: Check MCP server status

1. Run the command to manage MCP Server Configuration:

    /mcp show
   

2. You should see miro-mcp listed and marked as enabled.

    MCP Servers
       
     ❯ ✓ miro                http , OAuth: authenticated  https://mcp.miro.com/
       ✓ atlassian-rovo-mcp  http , OAuth: authenticated  https://mcp.atlassian.com/v1/mcp/authv2
   
    Built-in:
    ✓ github-mcp-server   http   https://api.business.githubcopilot.com/mcp/readonly
       
    Config:
    /Users/ivan/.copilot/mcp-config.json
   

Configure Claude Code with Miro MCP Server

You can connect the same Miro MCP Server to Claude Code, Anthropic’s agentic coding tool. Claude Code supports remote HTTP MCP servers with OAuth out of the box.

Step 1: Add the Miro MCP Server

Run the claude mcp add command with the HTTP transport:

claude mcp add --transport http miro https://mcp.miro.com/

By default the server is added to the current project (local scope). Use --scope user to make it available across all your projects:

claude mcp add --transport http --scope user miro https://mcp.miro.com/

Step 2: Authenticate via OAuth

After adding the server, claude mcp list shows it with ! Needs authentication. That’s expected — the next step completes the sign-in.

1. Start Claude Code and open the MCP panel:

    /mcp
   

2. Select miro from the list, press Enter, and choose Authenticate.
3. Your browser opens Miro’s authorization page. Select your team, review the requested permissions, and approve.
4. Back in Claude Code, the server’s status changes to connected.

Claude Code stores the access token securely and refreshes it automatically. If a refresh fails, Claude Code points you to /mcp where you can Re-authenticate.

Step 3: Check MCP server status

1. Open the MCP panel to manage your servers:

    /mcp
   

2. You should see miro listed and marked as connected, alongside any other configured servers.

    Manage MCP servers
    4 servers
   
      Project    (/Users/ivan/git/ibanFR/ibanfr.github.io/_posts/.mcp.json)
      MCPs
    ❯ atlassian-rovo-mcp · ✔ connected · 31 tools
      claude-code-docs · ✔ connected · 3 tools
      github · ✔ connected · 47 tools
   
      User MCPs (/Users/ivan/.claude.json)
      miro · ✔ connected · 35 tools
   

Test the connection

With the server connected, try a basic Miro-related prompt in your assistant:

can you list my miro boards?

Your assistant should respond with a list of your Miro boards, confirming the connection is working. You can recheck the server status at any time — /mcp show in Copilot CLI, or /mcp and claude mcp list in Claude Code.

Security and best practices

• Token expiry — OAuth tokens are managed automatically by Miro. If prompted to re-authenticate, follow the OAuth flow again
• Permissions — Only grant access to teams whose boards you want your assistant to interact with
• Audit — Check your Miro workspace’s audit log to see which connected apps have accessed your boards
• Revoke access — If you no longer want a client to access Miro, remove the MCP server from its configuration (e.g. ~/.copilot/mcp-config.json for Copilot CLI, or claude mcp remove miro for Claude Code)

Other resources

• Miro MCP Official Docs
• GitHub Copilot CLI Documentation
• Claude Code MCP Documentation
• Model Context Protocol Overview
• OAuth 2.1 Specification