READMEmain

Tableau MCP Server

A Model Context Protocol (MCP) server for Tableau Server and Tableau Cloud, built on the official mcp Python SDK (FastMCP). It authenticates via a Personal Access Token (PAT) and exposes read/write tools backed by the Tableau REST API.

Features

PAT-based sign-in with automatic sign-out on shutdown (via FastMCP lifespan)
Async httpx client with typed exceptions (TableauError, AuthenticationError, APIError, NotFoundError)
Tools for sites, projects, workbooks, views, data sources
View data (CSV) and image (PNG) retrieval
Extract refresh and workbook download helpers

Installation

git clone https://github.com/asklokesh/tableau-mcp-server.git
cd tableau-mcp-server
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

Configuration

All settings are read from environment variables with the TABLEAU_ prefix (a .env file next to the working directory is also supported).

Variable	Default	Description
`TABLEAU_SERVER_URL`	(required)	Base URL, e.g. `https://10ax.online.tableau.com` or `https://tableau.example.com`
`TABLEAU_API_VERSION`	`3.22`	Tableau REST API version
`TABLEAU_SITE_CONTENT_URL`	`""`	Site content URL; empty string targets the default site
`TABLEAU_PAT_NAME`	(required)	Personal Access Token name
`TABLEAU_PAT_SECRET`	(required)	Personal Access Token secret
`TABLEAU_TIMEOUT`	`60`	HTTP request timeout (seconds)

Running

tableau-mcp
# or
python -m tableau_mcp.server

Claude Desktop configuration

{
  "mcpServers": {
    "tableau": {
      "command": "tableau-mcp",
      "env": {
        "TABLEAU_SERVER_URL": "https://10ax.online.tableau.com",
        "TABLEAU_SITE_CONTENT_URL": "mysite",
        "TABLEAU_PAT_NAME": "my-pat",
        "TABLEAU_PAT_SECRET": "xxxxxxxxxxxx"
      }
    }
  }
}

Tools

Tool	Description
`list_sites`	List all sites reachable with the current PAT
`list_projects`	List projects on a site
`list_workbooks`	List workbooks, optionally filtered by project name
`list_views`	List views on a site or for a specific workbook
`list_datasources`	List published data sources
`query_view_data`	Return view data as CSV (optional `vf_*` filters)
`query_view_image`	Return view rendering as a base64-encoded PNG
`refresh_datasource`	Trigger an extract refresh for a data source
`refresh_workbook`	Trigger an extract refresh for a workbook
`download_workbook`	Download a workbook as base64-encoded `.twb`/`.twbx`

Development

pip install -e ".[dev]"
pytest -x --tb=short
ruff check src tests

License

MIT License - see LICENSE file for details.