Data Engineering
Security scan passedEquips your coding agent with a data engineer, pipeline commands and skills for Airflow, dbt, Spark and data quality.

Lets your agent search, query and chart the data in your own Metabase, and build questions and dashboards with your permissions.
by metabase· Data & analyticsUnclaimed · Yours? Claim it
4 images
About
Metabase's own MCP server, built into Metabase itself. Your agent connects straight to your Metabase and can search your content, browse your databases, run queries, draw charts, and create questions, dashboards and alerts, all limited to what your Metabase account is allowed to see and do. What you can do with it - Search questions, dashboards, models, metrics and documents, and browse collections - Browse databases, schemas, tables and fields, and look up business terms in your glossary - Answer questions from your data with structured queries, or with raw SQL when you allow it - Show results as an interactive chart in agents that can display one, with an Explore in Metabase button - Create and update questions, dashboards, documents, collections, metrics, segments and measures - Set up alerts and dashboard subscriptions that email or Slack your data How it connects There's nothing to install: the server runs inside your Metabase. Once an admin turns it on in Admin > AI > MCP, add your Metabase's address followed by /api/metabase-mcp to your agent, for example https://acme.metabaseapp.com/api/metabase-mcp. Admins can copy the exact address from Admin > AI > MCP > Settings. You sign in to your own Metabase with OAuth and approve the connection; there's no API key to paste. Choose what it can do The consent screen lists permissions for reading content, running queries, writing content, running raw SQL, and creating alerts and subscriptions. Reading and running queries are always included; untick the rest and it can only read and query. No permission gives it more than your Metabase account already has. Good to know The MCP server needs Metabase's AI features turned on, but no AI provider: your agent brings the model. Inline charts show in agents that can display them, such as Claude and ChatGPT, once an admin switches them on in Metabase. On Pro and Enterprise plans, admins can choose which tools each group can use and see MCP usage analytics.
Once it’s added
“Show me last month's signups by plan as a chart.”
“Find our revenue dashboard and sum up what it shows.”
Before you add it
Reads only
Your questions, dashboards, models and collections · Your databases' tables and fields · Query results, as rows or an inline chart
Makes changes · your agent can ask first
Creates and updates questions, dashboards, documents and collections · Creates metrics, segments and measures · Sets up alerts and dashboard subscriptions
From people who use it
Step by step
Nothing new reaches your agent until a person has looked at it.
Plugins people install alongside Metabase MCP.
No reviews yet. Used Metabase MCP? Say how it went.
Write a reviewPlugins people install alongside Metabase MCP.
Equips your coding agent with a data engineer, pipeline commands and skills for Airflow, dbt, Spark and data quality.
Lets your agent query your PostHog analytics, triage errors, and create feature flags and experiments from plain-language asks.
Data analytics skills for product managers: SQL from plain English, cohort retention analysis and A/B test reads with a ship-or-stop call.
Nobody has shared how they use Metabase MCP yet.
“What does 'active customer' mean in our Metabase?”
“Save that as a question and add it to the Growth dashboard.”
“Alert me when daily orders drop below 100.”
“Email the Sales dashboard to the team every Monday morning.”
Scanned on every change, and again when our rules improve.
Tell us and a person on the trust team will look at it. Or email [email protected].
Security scan
Passed
Safe to install in every agent it supports. Runs remotely on {metabase_host} · not an mcp server.