Guides

Kanban Studio MCP

Kanban Studio MCP Overview

Understand how AI agents connect to your Kanban Studio workspace through MCP.

Author: Ethan Navarro (Dev). Reviewed by the Kanban Studio team.

Kanban Studio MCP follows the open Model Context Protocol. If you are new to the protocol itself, review the official specification for transport and tool semantics. MCP specification.

This guide focuses on practical online workflow execution: connecting an MCP server, managing kanban boards, and controlling board actions with scoped credentials.

What is Kanban Studio MCP?

Kanban Studio MCP is an MCP server endpoint that lets AI agents read and update workspace data through a controlled tools interface.

It is designed for practical automation workflows where your agent can work with boards, lists, cards, and comments.

In practice, this gives AI assistants a controlled interface to work with boards, lists, and cards while keeping all permission checks in your backend.

How it works

Your MCP client calls the Kanban Studio MCP endpoint, the server authenticates with your API key, checks scopes, and executes authorized Convex functions.

A typical request flow is: MCP client -> Kanban Studio MCP route -> OAuth bearer validation -> scope verification -> Convex authorization by board role -> audited response. This layered model reduces accidental data exposure across workspaces.

json
{
  "mcpServers": {
    "kanbanstudio-local": {
      "transport": "streamable_http",
      "url": "https://kanbanstudio.site/api/mcp"
    }
  }
}

Security model

  • API keys are scoped and can be revoked or rotated.
  • Final authorization is enforced per board/workspace membership and role.
  • MCP calls are rate-limited and audited.

For production usage, prefer OAuth clients per integration when possible. If you must use API keys for a legacy client, keep them isolated and rotate regularly.

You should also keep scopes as narrow as possible and avoid sharing keys across teams or environments.

Next steps

From here, continue with the connection guide to configure your client and verify tool execution. Then review supported tools to choose minimal scopes for each automation.

Frequently asked questions

How do I manage kanban boards with an MCP server?

Connect your MCP client to the Kanban Studio MCP endpoint, complete OAuth, and run read and write tools with board-level authorization.

Can I still use API keys for board operations?

Yes, as a legacy fallback. OAuth is recommended by default and API keys should only be used when your MCP client cannot complete OAuth.

Does this work for online project management workflows?

Yes. The MCP route is available over HTTP and is designed for online workflows where AI agents manage boards, lists, and cards.

How do I verify my MCP setup is working?

Run a read tool such as list_boards from your MCP client. If boards are returned, authentication and board permissions are valid.