Tableau MCP

Tableau MCP is an MCP server that provides tools, resources, and prompts for AI applications interacting with Tableau Server and Tableau Cloud. Developed for BI developers, data analysts, and enterprise engineers, it connects LLM clients directly to Tableau environments. The server allows language models to run natural language queries across published data sources, inspect workbook metadata, retrieve views, and discover content stored inside a Tableau site. By exposing Tableau's analytical capabilities through the Model Context Protocol, teams can retrieve key business metrics, pull rendered dashboard views, and explore content engagement metrics without manually navigating the Tableau web interface. It supports both a hosted managed endpoint using OAuth 2.1 authentication for Tableau Cloud and self-hosted deployments via npm, Docker, or Heroku for on-premise Tableau Server installations requiring Personal Access Token authentication.

Category: CRM, ERP & E-commerce

Tags: bi, business-intelligence, data-visualization, tableau

Visit Tableau MCP

How to install and configure Tableau MCP

1. Ensure you have Node.js 22.7.5 or later installed on your system. 2. Open your MCP client configuration file (such as Claude Desktop's claude_desktop_config.json). 3. Add the Tableau server entry under the mcpServers object: 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" } } } } 4. Replace SERVER, SITE_NAME, PAT_NAME, and PAT_VALUE with your actual Tableau deployment URL and credentials. 5. Save the configuration and restart your MCP client.

What you can do with Tableau MCP

• Querying published data sources like Superstore in natural language to extract summary metrics and rankings for specific date ranges. - Finding content by querying metadata, such as identifying the most viewed workbooks and their ownership details over the past year. - Retrieving and rendering visual dashboard views directly into the chat interface by specifying workbook and project names. - Automating business intelligence inquiries within team clients like Slack or Claude without manual navigation through Tableau Server. - Enforcing per-user permissions and secure access control dynamically using Tableau Cloud OAuth 2.1 integrations.

Key facts

• https://github.com/tableau/tableau-mcp
• CRM, ERP & E-commerce, Data & Analytics
• bi, business-intelligence, data-visualization, tableau

Part of MCP Servers

How do I install Tableau MCP?

You can run Tableau MCP locally without manual building by using npx. In your MCP client configuration file, configure the tableau server command as npx with arguments -y and @tableau/mcp-server@latest. Provide the required environment variables including SERVER, SITE_NAME, PAT_NAME, and PAT_VALUE. Node.js version 22.7.5 or later is required.

What can Tableau MCP do?

Tableau MCP lets AI agents query published Tableau data sources, explore metadata to find workbooks and views, and retrieve dashboard visual elements. Users can ask natural language questions about metrics, view analytics usage stats, and pull specific charts directly into chat clients.

Which MCP clients work with Tableau MCP?

Tableau MCP works with common MCP-compatible clients including Claude Desktop, Slack, and ChatGPT. For Tableau Cloud, clients can connect directly to the hosted managed service at mcp.tableau.com using OAuth 2.1 authentication, or connect locally to self-hosted instances.

Does Tableau MCP support Tableau Server and Tableau Cloud?

Yes, Tableau MCP supports both environments. Tableau Cloud users can use the hosted service with OAuth 2.1 or self-host it. Tableau Server customers can self-host using npx, Docker, Heroku, or by building from source using Personal Access Tokens for authentication.