Gene Library Courses Download Pricing Contact Sign in
monday logo
productivity Official Website

monday.com MCP Server – Work OS Automation

monday.com's official MCP servers connect AI assistants to monday.com workspaces and development tools. Use the hosted MCP server for secure workspace access to boards, items, docs, reporting, CRM, and operational workflows, or use the local package for custom API and app-framework development.

#work-management#project-tracking#crm

Overview

monday MCP connects AI assistants to monday.com so they can understand and act
on structured work data. The hosted server gives supported AI clients secure
access to a monday.com workspace through OAuth, while the official npm package
can run locally for API-based workflows and app-framework development.

What the MCP server enables

Hosted monday MCP is designed for workspace automation and reporting. Depending
on account permissions, the AI Connector settings, and user approval, an agent
can:

  • Retrieve board, item, workspace, folder, form, and doc context.
  • Generate project, sprint, deadline, blocker, and executive summaries.
  • Convert notes or transcripts into structured items.
  • Assign owners, update statuses, adjust dates, and add item updates.
  • Support CRM workflows such as creating leads, updating deals, and logging next
    steps.
  • Create or attach monday docs for specs, SOPs, retros, and project notes.
  • Build dashboards and widgets or answer questions using board insights.
  • Run multi-step operational flows such as creating bugs, incidents, support
    requests, and checklists.

monday also documents a separate apps-framework MCP mode. In apps mode, an AI
assistant can list and create development apps, create app features, manage app
versions, promote versions to live, check deployment status, set environment
variables, search app storage, export storage data, and retrieve development
context from official apps documentation.

When to use it

Use hosted monday MCP when an AI assistant should help teams operate work inside
monday.com: summarizing projects, preparing sprint reports, cleaning up tasks,
updating CRM pipelines, or finding cross-board blockers. Use local stdio when a
developer wants a local MCP process, a specific API token, dynamic API tools, or
access to monday's app-development mode.

Connection and authentication

The recommended hosted endpoint is https://mcp.monday.com/mcp. Hosted MCP is
preinstalled on monday.com accounts and uses OAuth consent when prompted by the
supported AI client. Admins may need to check Admin → Permissions → AI
Connectors if users cannot connect.

Local stdio uses npx @mondaydotcomorg/monday-api-mcp@latest with
MONDAY_TOKEN. The apps framework mode uses the same package with --mode apps, or with -t to pass a token as documented. Node.js v20 or higher and
npm v5.2.0 or higher are documented prerequisites.

Key considerations

monday MCP can create and update real workspace data, so use human review before
applying status changes, assigning owners, updating CRM deals, creating docs, or
running multi-step workflows. Access is governed by the user's account
permissions, OAuth authorization, workspace controls, and AI Connector settings.
Local tokens should be scoped and stored securely. Dynamic API Tools are
documented as beta and disabled by default. The apps-framework mode can promote
app versions and set environment variables, so restrict it to trusted
developers and review changes before moving versions live.

Supported Transports

streamable_http

URL: https://mcp.monday.com/mcp

streamable_http

URL: https://mcp.monday.com/mcp

stdio

Command: npx

Args:

  • @mondaydotcomorg/monday-api-mcp@latest

stdio

Command: npx

Args:

  • @mondaydotcomorg/monday-api-mcp
  • -t
  • ${MONDAY_TOKEN}
  • --mode
  • apps

Frequently Asked Questions

When should an AI agent use the monday.com MCP server?
Use it when an agent needs to retrieve or update work in monday.com, such as summarizing projects, creating tasks from meeting notes, updating statuses and owners, reporting blockers, managing CRM next steps, creating docs, or running approved operational workflows.
What does monday.com MCP add to an AI agent's capabilities?
It gives the agent live, permission-aware access to monday.com workspace data and tools, allowing the agent to work with boards, items, docs, dashboards, CRM data, reports, and app-development resources rather than relying on copied exports or static model knowledge.
What can an AI agent access or manage through monday.com MCP?
Hosted MCP can work with accessible boards, items, docs, dashboards, workspace discovery, CRM workflows, project reporting, and operational task updates. Apps mode can manage development apps, features, versions, deployments, environment variables, and app storage according to the API token's authorization.
How is authentication configured for monday.com MCP?
Hosted MCP uses OAuth consent through the supported AI client and may be controlled by account Admin → Permissions → AI Connectors settings. Local stdio uses a monday.com API token supplied as MONDAY_TOKEN or with the documented -t argument. Store tokens securely and avoid committing them.
Which transport should be used for monday.com MCP?
Use the hosted Streamable HTTP endpoint at https://mcp.monday.com/mcp for most workspace automation because monday.com recommends it. Use local stdio with @mondaydotcomorg/monday-api-mcp when a local process, custom token, dynamic API tools, or apps-framework mode is required.