Tableau MCP Server

Official

Tableau's official MCP Server, helping agents see and understand data.

How to install ↓ Compare GitHub source repository ↗

Category: Database #25 of 58
Stars: ★ 363 Very popular
Transport: stdio (local process) · Streamable HTTP
Runtime: Node.js 22.7.5+ · Hosted (no local runtime) · Docker
Credentials: API key / credential required
License: Apache-2.0
Last commit: 1 day ago

57FMRS · C

Tableau MCP is a Tableau Supported, Apache-2.0 licensed MCP server that offers both a hosted cloud service and self-hosted deployment, with authentication via OAuth 2.1 or PAT credentials in environment variables, and it honors existing per-user Tableau permissions. It suits users and developers already on Tableau who want AI agents to query data, explore content, and inspect views, but it is not applicable without a Tableau environment or when agent interaction is unnecessary. The main risks are credential handling and data exposure scope; confirm permission boundaries before deploying and use the official docs to diagnose common issues.

Strongest · Maintenance 14/20 Weakest · Reliability 8/20

Reliability

8/20

Security and permissions

10/20

Maintenance

14/20

Documentation

13/20

Setup experience

12/20

Why each score

Reliability 8/20: This is a static review with only the README and repository metadata; no SERVER MANIFEST was cached and no source files or actual CI workflow content were available (the README only links a CI badge, which is not execution evidence). The README describes both a hosted service and an npx self-hosted path, so the happy path is plausible, but nothing verifies that MCP initialization and tool-list handshake succeed cleanly, that declared tools match real behavior, or that dependencies and errors are controllable. Per the static calibration rule, reliability cannot exceed 12 without verifiable execution evidence, and unknown items are deducted rather than assumed good, hence 8.
Security and permissions 10/20: The hosted endpoint mcp.tableau.com uses OAuth 2.1, requires each user to sign in with their own Tableau Cloud identity, and reuses existing per-user permissions — a genuine least-privilege signal. However, the self-hosted example places PAT_NAME and PAT_VALUE as plaintext env values in client configuration, i.e., long-lived credentials written in cleartext, with no guidance on rotation, scope minimization, or secure storage alternatives. The README also does not disclose data-flow boundaries or whether side-effecting operations (publish/delete content) require explicit confirmation. No red line is crossed (no credential theft, covert exfiltration, or real leaked tokens), so it is not placed in the 0-4 band, but permission and confirmation disclosure is incomplete: 10.
Maintenance 14/20: Positive signals: an official Tableau-supported badge, Apache-2.0 license, a published npm package (@tableau/mcp-server), a CI workflow referenced by badge, 339 stars, and a maintained documentation site indicate clear organizational ownership and a release channel. Deductions: 83 open issues is relatively high, and there is no evidence of release notes, commit cadence, or issue response times, so dependency updates and a security-response channel cannot be assessed. Active but with opaque governance and versioning: 14.
Documentation 13/20: The README gives an overview, two install paths (hosted and npx), configuration JSON examples, several example prompts with sample responses, and links to the official docs site plus enterprise/self-hosted guides — a clear layered structure. Deductions: visible in-repo content omits tool parameter lists, call limits, quota/cost information, troubleshooting, and permission mapping details; the required environment variables (SERVER, SITE_NAME, PAT_*) appear only as an example with no field semantics or scope notes. Usable but with hidden assumptions: 13.
Setup experience 12/20: The install path is explicit: hosted use requires pointing at https://mcp.tableau.com and completing OAuth; local use is a single npx command, with the Node.js 22.7.5+ requirement, a complete mcpServers JSON snippet, and entry points for Docker, source builds, and one-click Heroku deployment, including client config examples. Deductions: no cached manifest and no end-to-end connectivity evidence; PAT credentials must be prepared by the user via external documentation, and the first real connection may be fragile across environments. Capped at 15 by the static calibration rule and reduced for uncertainty: 12.

Static review · not runListed 2026-09-18

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkConnects to a database

Best for

Teams on Tableau Cloud that want a quick managed OAuth-based setup
Tableau Server customers who need self-hosted infrastructure
Developers building AI applications that integrate with Tableau

Not for

Users without a Tableau Cloud or Tableau Server environment or license
Scenarios that only need static report exports without agent interaction
Local environments that cannot meet the Node.js 22.7.5 or later requirement (self-hosted option)

Required permissions

Access to the Tableau site and data (configured for self-hosting via SERVER, SITE_NAME, PAT_NAME, PAT_VALUE)
OAuth 2.1 sign-in with the user's own Tableau Cloud identity under the hosted service, inheriting existing per-user permissions
Read access to Tableau content to support data queries, content exploration, and view inspection

Risks and side effects

Leaked PAT credentials (PAT_NAME, PAT_VALUE) could allow unauthorized Tableau access, so they must be kept secure
The data an agent can read is bounded by the signed-in identity's permissions, but sensitive data the user can access may still be exposed
The self-hosted option requires maintaining your own runtime and security updates
Using the hosted service means Tableau content is processed through https://mcp.tableau.com

Setup

Before you start

Runtime：Node.js 22.7.5+ · Hosted (no local runtime) · Docker

SERVER required URL of your Tableau server, e.g. https://my-tableau-server.com.

PAT_NAME required Name of your Tableau Personal Access Token, created in your Tableau account settings.

PAT_VALUE requiredsecret Secret value of your Tableau Personal Access Token, generated in your Tableau account settings.

Other optional settings (1)

SITE_NAME optional Tableau site name, needed for multi-site instances.

Hosted (recommended for Tableau Cloud): point any MCP-compatible client at https://mcp.tableau.com and complete the OAuth sign-in flow when prompted. Self-hosted/local: install Node.js 22.7.5 or later, add npx -y @tableau/mcp-server@latest to your MCP client configuration, and set the SERVER, SITE_NAME, PAT_NAME, and PAT_VALUE environment variables. See the official Getting Started guide for Docker, building from source, and other self-hosted options; a Heroku deploy button is also available.

MCP config · JSON

{
  "mcpServers": {
    "tableau": {
      "command": "npx",
      "args": [
        "-y",
        "@tableau/mcp-server@latest"
      ],
      "env": {
        "SERVER": "https://my-tableau-server.com",
        "SITE_NAME": "my_site",
        "PAT_NAME": "my_pat",
        "PAT_VALUE": "pat_value"
      }
    }
  }
}

Works in clients that read an "mcpServers" key. For clients such as VS Code that use a different key, the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "tableau": {
      "command": "npx",
      "args": [
        "-y",
        "@tableau/mcp-server@latest"
      ],
      "env": {
        "SERVER": "https://my-tableau-server.com",
        "SITE_NAME": "my_site",
        "PAT_NAME": "my_pat",
        "PAT_VALUE": "pat_value"
      }
    }
  }
}

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

Terminal

claude mcp add tableau -e SERVER=https://my-tableau-server.com -e SITE_NAME=my_site -e PAT_NAME=my_pat -e PAT_VALUE=pat_value -- npx -y @tableau/mcp-server@latest

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

Generate and validate client config

Check that it works

After configuring the tableau server in your MCP client, ask a data question such as 'For the Superstore Datasource, what are the top 5 states with the most sales in 2025?' — a data-backed answer confirms the connection works.

Troubleshooting

Verify that Node.js is version 22.7.5 or later
Check that SERVER, SITE_NAME, PAT_NAME, and PAT_VALUE are set correctly
When using the hosted service, confirm the OAuth sign-in flow was completed
Tableau Server customers should follow the Enterprise Deployment and Self-Hosted Getting Started guides rather than the hosted service docs
If the connection still fails, consult the official documentation at https://tableau.github.io/tableau-mcp/ and the client integration guides

Things to try

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

For the Superstore Datasource, what are the top 5 states with the most sales in 2025?
Find me the most viewed workbook within the last year.
Show me the Economy view in the Finances project.

Use cases

Ask an AI client questions about Tableau datasources, e.g. the top 5 states by sales in 2025

Find the most viewed workbook within the last year

Show a specific view inside a project

Bring Tableau data insights into AI clients such as Slack, Claude, and ChatGPT

Supported clients

Claude

ChatGPT

Slack

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

Overview

Tableau MCP is a suite of developer primitives — tools, resources, and prompts — that makes it easier for developers to build AI applications that integrate with Tableau. It is offered both as a hosted service at https://mcp.tableau.com, which uses OAuth 2.1 so every user signs in with their own Tableau Cloud identity while existing per-user permissions are enforced automatically, and as a self-hosted/local option run via npx @tableau/mcp-server@latest, which requires Node.js 22.7.5 or later and uses environment variables such as SERVER, SITE_NAME, PAT_NAME, and PAT_VALUE to connect to a Tableau Server or site. It can also be deployed to Heroku and works with common MCP clients such as Slack, Claude, and ChatGPT for querying data, exploring content, and viewing specific views.

Similar servers

MCP ClickHouse 79 · B

Connect ClickHouse to your AI assistants

★ 900 · Tools 4 Compare with this →

ChEMBL MCP Server 75 · B

Unofficial MCP server providing advanced access to the ChEMBL chemical database

★ 89 · Tools 27 Compare with this →

MCP Toolbox for Databases 74 · B

Google's official database MCP toolbox — define AI-accessible database tools via config

★ 16.8k Compare with this →

DBHub 74 · B

Bytebase's official, minimalist two-tool database MCP server

★ 3.8k · Tools 2 Compare with this →

Compare with other MCP servers

Source revision 86b058f33692 Data synced 2026-11-02 Read the FMRS scoring method