monday.com MCP Server

Official

Connect AI agents securely to the monday.com Work OS

How to install ↓ Compare GitHub source repository ↗

Category: Collaboration #25 of 55
Stars: ★ 428 Very popular
Transport: stdio (local process) · SSE · Streamable HTTP
Runtime: Node.js 20+ · Hosted (no local runtime)
Credentials: API key / credential required
License: MIT
Last commit: today
Tools: 17

55FMRS · C

An officially maintained monday.com MCP server offering both a hosted remote service and a local npx option, covering core board/item/column/form operations plus a beta dynamic GraphQL toolset; users should safeguard API tokens and be mindful of the write-access risks.

Strongest · Documentation 15/20 Weakest · Security and permissions 4/20

Reliability

9/20

Security and permissions

4/20

Maintenance

13/20

Documentation

15/20

Setup experience

14/20

Why each score

Reliability 9/20: server.json declares both streamable-http and sse remote transports, and the README's tool table (create_item, delete_item, get_board_schema, etc.) is internally consistent with the described package structure, giving a plausible baseline. However, this is a static-only review, and no CI workflow or committed tests covering key paths were present in the supplied material, so the init/tool-list handshake and declared-vs-actual tool behavior cannot be confirmed, nor can error handling controllability. Per the static calibration rule, reliability is capped at 12 absent verifiable execution evidence; scored 9, deducted for missing test/CI evidence and undocumented error-handling behavior.
Security and permissions 4/20: By default (i.e., outside read-only mode) the server exposes irreversible-effect tools — create_item, delete_item, delete_column — and neither the README nor the manifest describes any confirmation step beyond whatever the MCP client itself may provide. The manifest also marks the Authorization header as isRequired:false, meaning authentication is optional for a remote service that can perform destructive writes, which is an unusual and risky default. While the full Dynamic API tool surface is disabled by default (a positive least-privilege signal) and no hardcoded secrets or covert exfiltration are evident, shipping default-enabled irreversible delete operations without a documented confirmation flow triggers the red-line condition, capping security at 0-4; scored 4.
Maintenance 13/20: The repository is maintained by monday.com's own team, uses the MIT license, is not archived, and shows community activity (417 stars, 43 open issues) with clear ownership attribution in the README. However, the supplied material contains no verifiable release history, changelog, issue-response SLAs, or dependency-update cadence, and the server.json version is still 0.0.1, suggesting an early-stage governance/versioning process. Scored 13.
Documentation 15/20: The README is well-layered, covering hosted vs. local install paths, multiple client configuration examples (Claude Desktop, Cursor, Gemini CLI), a full tool table with descriptions, authentication instructions, a security best-practices section, and explicit disclosure of the Dynamic API tools' beta status and limitations (e.g., incompatibility with read-only mode); no real secrets appear in examples. It lacks rate-limit/cost/quota disclosure and a dedicated troubleshooting section, pointing only to GitHub Issues for support. Scored 15.
Setup experience 14/20: Local installation is a single npx command plus an environment variable, and the hosted MCP option requires only a URL; configuration examples span several mainstream clients with low apparent complexity. Since this is a static-only review with no CI or integration-test evidence in the supplied material confirming these install paths actually work end-to-end, setup is capped at 15 per the static calibration rule; scored 14 given the completeness of the documented examples.

Static review · not runListed 2026-08-26

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

Teams already using monday.com for project/task management who want AI-driven board operations
Developers connecting AI agents to structured work data for task automation
Users who prefer the officially hosted service over local setup

Not for

Teams not using the monday.com platform
Fully offline environments (even local mode requires network access to the monday.com API)
Cases needing read-only guarantees without explicitly enabling read-only mode

Required permissions

Requires a monday.com API token (personal access token or OAuth) to act on the user's account
Default mode grants read/write access — can create, modify, and delete boards/items/columns, so token scope should be handled carefully
Dynamic API Tools require full API access and are incompatible with read-only mode

Risks and side effects

A leaked API token exposes read/write access to all boards under that account
Write tools such as delete_item and delete_column can cause irreversible data loss
Dynamic API Tools allow arbitrary GraphQL queries/mutations, widening the potential blast radius of misuse

Setup

Before you start

Runtime：Node.js 20+ · Hosted (no local runtime)

MONDAY_TOKEN requiredsecret Your monday.com personal API token, generated in monday.com under your avatar → Developers → My access tokens. Required for authentication.

The recommended path is the official hosted MCP: add the URL https://mcp.monday.com/mcp (streamable-http) or https://mcp.monday.com/sse (SSE) directly in your client's MCP settings, supplying a monday.com API token via the Authorization header (as 'Bearer your_token' or the bare token; optional but required for authenticated access). For local use, configure your client to run npx @mondaydotcomorg/monday-api-mcp@latest with the MONDAY_TOKEN environment variable set to a personal access token (generated from your monday.com avatar menu → Developers → My access tokens). Gemini CLI users can install via gemini extensions install or gemini mcp add, then run /mcp auth monday to complete OAuth. Requires Node.js v20+ and npm v5.2.0+.

claude_desktop_config.json

{"mcpServers": {"monday-api-mcp": {"command": "npx", "args": ["@mondaydotcomorg/monday-api-mcp@latest"], "env": {"MONDAY_TOKEN": "your_monday_api_token"}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "monday-api-mcp": {
      "command": "npx",
      "args": [
        "@mondaydotcomorg/monday-api-mcp@latest"
      ],
      "env": {
        "MONDAY_TOKEN": "your_monday_api_token"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add monday-api-mcp -e MONDAY_TOKEN=your_monday_api_token -- npx @mondaydotcomorg/monday-api-mcp@latest

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After adding the config, restart your client and confirm the monday tools (e.g. get_board_schema, create_item) appear in the tool list; then ask "What items do I have in board 123?" — a correct answer proves the connection works.

Troubleshooting

Verify MONDAY_TOKEN is set (or passed via --token) with a valid API token
For local npx setup, confirm Node.js v20+ is installed
If expected tools are missing, check the --mode (default/apps) and --enable-dynamic-api-tools flags
For hosted connection failures, verify the Authorization header format ('Bearer your_token' or bare token)

Things to try

Once connected, you can ask your AI assistant things like:

What items do I have in board 123?
Can you create a board to manage my project?
Search board 123 for items named "Update the UI" and add an update to them
Create a new status column called 'In Review' on board 123

Tools 17

create_item writes

Create a new item in a monday.com board with specified column values

delete_item destructive

Delete an item from a board permanently

get_board_items_by_name read-only

Search for items by board ID and term/name

create_update writes

Add an update/comment to a specific item

change_item_column_values writes

Modify the column values of an existing item

move_item_to_group writes

Move an item to a different group within the same board

create_board writes

Create a new monday.com board with specified columns

get_board_schema read-only

Retrieve the structure of columns and groups for a board

Show 9 more tools

create_group writes

Create a new group in a monday.com board

create_column writes

Add a new column to an existing board

delete_column destructive

Remove a column from a board

list_users_and_teams read-only

Retrieve user or team details by id, name, or by searching the account

create_form writes

Create a new monday.com form

get_form read-only

Get a form by its token

all_monday_api writes

Generate and execute any GraphQL query or mutation dynamically

get_graphql_schema read-only

Fetch monday.com's GraphQL schema to understand available operations

get_type_details read-only

Retrieve detailed information about specific GraphQL types

Use cases

Let an AI assistant create, query, and update items and column values on monday.com boards

Automate project management workflows across teams

Use Dynamic API Tools to build custom GraphQL queries for cross-board reporting

Supported clients

Claude Desktop

Cursor

Gemini CLI

ChatGPTPartial support

Microsoft Copilot StudioPartial support

Listed from the project's documentation, not tested by this site.

Overview

The monday.com MCP server, maintained by the monday.com AI team, lets AI agents read and act on monday.com boards, items, columns, groups, and docs. It's offered both as an official hosted remote service (streamable-http and SSE) and as a local stdio server run via npx, working with clients like Claude Desktop, Cursor, and Gemini CLI. Beyond the built-in tools, a beta 'Dynamic API Tools' set can generate and execute arbitrary GraphQL queries against the full monday.com API.

Similar servers

Atlassian Rovo MCP Server 66 · C

Securely connect Atlassian Cloud to AI tools.

★ 1.1k Compare with this →

Notion MCP Server 64 · C

Official Notion MCP server for querying, editing, and managing Notion pages and databases via AI agents.

★ 4.7k · Tools 9 Compare with this →

WhatsApp Channel for Claude Code 64 · C

Drive your Claude Code session straight from WhatsApp

★ 104 · Tools 10 Compare with this →

Anytype MCP Server 64 · C

Official MCP server that lets AI assistants manage your encrypted Anytype wiki in natural language.

★ 532 Compare with this →

Compare with other MCP servers

Source revision 78b176609cce Data synced 2026-10-12 Read the FMRS scoring method