MCP servers for to-do apps

monday.com MCP server

Yes. monday.com hosts an official MCP server at mcp.monday.com/mcp. Clients sign in with OAuth, Claude and ChatGPT are pre-approved, and every tool call runs as you, inside your monday.com permissions and any limits your admin has set.

Official serverServer URLhttps://mcp.monday.com/mcp

How to connect it

  1. Claude or ChatGPT: add monday.com from the client's connector or app directory and approve the OAuth consent screen.
  2. Cursor: add a server with the URL https://mcp.monday.com/mcp under mcpServers.
  3. Any other MCP client: point it at the same URL; discovery, registration and sign-in happen automatically. If the consent screen says Admin approval required, an account admin has to approve the connection first.
  4. Prefer a local server? monday.com also publishes npx @mondaydotcomorg/monday-api-mcp, which runs with an API token and has a --read-only flag.

What it can do

  • Create items, change column values, move items between groups and post updates
  • Create boards, groups and columns, and read a board's schema
  • List users and teams, and create or read WorkForms
  • Run arbitrary monday.com API queries through Dynamic API tools, which are in beta and off by default

What it doesn't do

  • The hosted connector is all-or-nothing: once it's enabled, anyone who can log in can connect from any MCP client. Tighter control means registering your own OAuth app with narrow scopes, such as boards:read.
  • A read-only switch exists on the local server, not as a documented option on the hosted one.
  • Dynamic API tools need full API access and don't work in read-only mode.
  • Its docs don't describe an approval step before an outside agent's change to an item counts.

Staying on monday.com?

If your board data matters, use your own OAuth app with only the scopes an agent needs. monday.com documents that those scopes cap what the MCP connection can do, even when the user has broader permissions.

When Done makes sense instead

Done is a to-do app built for handing work to an AI agent, with the parts an agent needs before you trust it with real tasks:

  • An approval gate: by default an agent can propose that a task is finished, but only you can close it. Done enforces that, so it holds even if an agent tries to skip it.
  • Questions instead of guesses: an agent parks a decision with its recommendation, and you answer in one click.
  • Plans before work: an agent can propose a plan that waits for your yes before it starts.
  • Lanes: name your agents and pin each task to one, so two agents don't pick up the same work.
  • Scoped access: tokens that are read-only or limited to one project or workspace, and budgets an agent can't raise.

Moving takes one step

monday.com exports a board to Excel from the board's ⋯ menu. Save the sheet as CSV and Done reads any CSV with a title column, or copy the item names and use Paste a list.

Done's import preview shows exactly what lands — projects, tasks, due dates, priorities — before you create an account, and nothing is uploaded.

Preview your monday.com import

Other to-do apps

Try Done with the agent you already use

Free to start. Connect Claude, ChatGPT or any MCP client in about a minute.

Sources

Information about monday.com current as of October 2026, checked against the sources above. MCP servers change quickly, so check them before you rely on a detail. monday.com is a trademark of its owner.