Skip to main content

Tableau MCP

Connect your AI tools to Tableau Cloud using the Model Context Protocol (MCP) through Cequence AI Gateway. The hosted Tableau MCP server lets AI assistants query your published data, find and read dashboards, publish workbooks, and ask questions about Tableau Pulse metrics. Each person signs in as themselves, so the assistant sees only what that person can already see in Tableau.

Overview​

Tableau MCP is a remote, vendor-hosted MCP server run by Tableau. You do not install anything; you connect to it from Cequence AI Gateway and each user signs in with their Tableau Cloud account.

  • Server URL: https://mcp.tableau.com
  • Transport: HTTP (Streamable HTTP)
  • Hosted by: Tableau (Salesforce)

The hosted server is for Tableau Cloud only. It is not available for Tableau Server. One URL serves every Tableau Cloud region; you choose your site when you sign in.

Supported authentication types​

TypeSupportedNotes
OAuth 2.1YesRecommended for Cequence AI Gateway. Each user signs in to their own Tableau Cloud site. You do not need to create an OAuth app in Tableau; Cequence AI Gateway identifies itself to Tableau automatically.
Personal access tokenNoOnly available when you run Tableau MCP yourself, not on the hosted server.

The MCP server always acts as the signed-in user, and every existing Tableau permission still applies.

What can you do with this MCP server​

The Tableau MCP server provides 26 tools, grouped below.

Data sources​

  • List the published data sources you can access
  • Read a data source's fields, data types, and parameters
  • Query a data source and get the results back as data

Workbooks and projects​

  • List workbooks and projects, and read a workbook's details
  • Download a workbook file
  • Publish a workbook to a project you choose

Views and dashboards​

  • List views and custom views
  • Get a view's underlying data, or an image of the view
  • Do the same for custom views that people have saved

Tableau Pulse​

  • List Pulse metric definitions and the metrics built from them
  • List the metrics you follow
  • Generate insights for a metric value, or an AI-written insight brief that answers a question about your metrics
  • Search across workbooks, views, data sources, projects, and more in a single query

Session controls​

  • Sign out of Tableau from the AI client, or reset the consent you gave so the next sign-in asks again

Prerequisites​

Before adding Tableau MCP in Cequence AI Gateway, ensure you have:

  • Access to Cequence AI Gateway at aigateway.cequence.ai
  • A Tableau Cloud account on the site you want AI tools to use
  • Tableau+ if you want Pulse insight briefs. Some metadata, such as field descriptions, requires Data Management.

What to ask your IT team​

Ask forWhy you need it
The Tableau Cloud site your team should connect toA connection is tied to one site
Access to the projects, workbooks, and data sources you needThe assistant can only reach what your account can see
Whether any Tableau MCP tools are turned off for the siteSite administrators can exclude groups of tools
Whether your plan includes Tableau+Pulse insight briefs need it

Example workflows​

  • "List the data sources in the Finance project."
  • "What were total sales by region last quarter in the Superstore data source?"
  • "Show me an image of the Executive Overview dashboard."
  • "Get the data behind the Pipeline by Stage view."
  • "Which Pulse metrics am I following, and what changed this week?"
  • "Why did weekly active users drop last week?"
  • "Publish this workbook to the Marketing project."

Connecting MCP server from Cequence AI Gateway​

  1. Log in to Cequence AI Gateway.
  2. Choose your tenant.
  3. Go to App catalogue.
  4. Filter by Remote MCP server.
  5. Search for Tableau and then select it.
  6. Select Create MCP server.
  7. Choose OAuth as the auth method. You do not need a client ID or secret from Tableau.
  8. Complete the setup as prompted, select tools, and deploy.
  9. When you first connect, sign in to Tableau Cloud, pick your site, and approve access.

Use the generated MCP server URL in your client as described in the Client Configuration docs. For detailed UI steps and screenshots, see Create a third-party MCP Server.

Additional information​

  • One site per connection: To switch to another Tableau Cloud site, disconnect and sign in again.
  • No stored data: The hosted server does not store your Tableau data. It passes each request to Tableau as the signed-in user.
  • Tableau Server: The hosted server does not support Tableau Server. Teams on Tableau Server run Tableau MCP themselves.
  • Timeout: The remote server uses a 30-second timeout for requests.
  • Official documentation: Hosted Tableau MCP, Client integrations.
Last updated on