Tableau MCP. Ask natural questions about your BI data.

You don't gotta click through dashboards to get answers anymore. This server hooks your AI client directly into Tableau Cloud or Server's metadata layer. Instead of just looking at what the rendered dashboard image shows you, your agent queries the underlying data structure—the actual workbooks and connections. It gives you instant access to deep BI insights without touching a GUI.

Browsing Site Assets and Workbooks
You need an overview? You can start by listing all project containers across the site using list_projects. That maps out the entire organizational hierarchy of your Tableau assets. From those projects, you'll find every available dashboard (or 'view') with list_views, giving you a master inventory of what exists.

To drill down into specific content, use list_workbooks to pull every workbook contained within a defined project container. Once you pinpoint the exact file, you run get_workbook; this fetches detailed metadata for that single workbook, letting your agent know its structure and contents.

Monitoring Data Health and Jobs
Data freshness is everything. You can check the health of all published data sources using list_datasources. This function doesn't just list them; it reports their last refresh date and connection status right away. If a source looks stale, you know where to poke around. For background tasks—like scheduled extract refreshes or complex ETL jobs—you monitor the entire site status with list_jobs.

This checks if those crucial processes finished successfully or if something failed, keeping your data pipeline running clean.

Auditing Governance and Users
Governance is a pain in the ass. You can audit who has access to what by querying user accounts using list_users. This function pulls a list of every active user account associated with the site, detailing their roles and group memberships. If you're trying to figure out permissions for an old employee or new hire, this is your go-to tool.

Putting It All Together
This server lets your agent perform complex audits in natural conversation. You can ask it to list all projects, then check the data source status of every published connection within those projects, and finally verify that the user accounts responsible for maintaining them still have active roles—all without you ever having to click a single button or run multiple reports yourself.

Your agent uses list_projects to get the container list. It employs list_views to inventory all dashboards available across every project. To see which workbooks are sitting inside, it hits list_workbooks. For deep dives on one file, get_workbook delivers the full metadata payload. If you need to verify data integrity, list_datasources tells you exactly when each dataset last updated.

You can monitor system reliability by running list_jobs, checking the status of background refresh tasks. Finally, for security and access control, list_users provides the definitive list of active accounts, their roles, and group affiliations on the entire site.