Metabase Ai Assistant
Drive Metabase dashboards, cards, and PostgreSQL-backed SQL from your agent with a large tool surface for product analytics and internal reporting.
Overview
Metabase AI Assistant is an MCP server for the Grow phase that provides 111 tools for Metabase SQL generation, dashboards, and PostgreSQL-backed analytics from your agent.
What is this MCP server?
- 111 MCP tools for Metabase administration, SQL generation, dashboards, and PostgreSQL-oriented workflows
- Authenticates with Metabase URL plus username/password or METABASE_API_KEY
- Optional direct PostgreSQL connection env vars for executing SQL outside Metabase UI
- Stdio npm package metabase-ai-assistant for local agent attachment
- Suited to iterating cards and collections as product questions change
- 111 MCP tools for SQL generation, dashboards, and PostgreSQL per server description
- Package version 3.2.0; npm identifier metabase-ai-assistant
- Stdio transport; requires METABASE_URL and Metabase auth credentials or API key
Community signal: 43 GitHub stars.
What problem does it solve?
You have Metabase and PostgreSQL metrics but waste hours clicking through collections or rewriting the same exploratory SQL for every product question.
Who is it for?
Technical founders with an existing Metabase instance who want agent-assisted reporting, card edits, and database exploration.
Skip if: Greenfield teams with no Metabase deployment or non-technical users who only need a static monthly PDF report.
What do I get? / Deliverables
Your agent can query Metabase, manage dashboard artifacts, and run scoped SQL so analytics loops happen in the IDE alongside code changes.
- Agent access to Metabase objects and query workflows across the published tool set
- Authenticated Metabase session via API key or user credentials
- Optional direct PostgreSQL query channel when DATABASE_* variables are configured
Recommended MCP Servers
Journey fit
Metabase is where founders answer “what is usage doing?” after launch, so Grow analytics is the primary shelf even though the same MCP helps wire integrations during Build. Analytics subphase covers SQL generation, dashboard management, and querying metrics—not deploying the Metabase server itself.
How it compares
Metabase admin and SQL MCP with 111 tools, not a lightweight read-only BI chat widget.
Common Questions / FAQ
Who is Metabase AI Assistant for?
Developers and solo builders who already use Metabase and PostgreSQL and want Cursor or Claude Code to manage analytics assets and run SQL safely.
When should I use Metabase AI Assistant?
Use it in Grow analytics when refining dashboards after launches, and during Build when wiring an internal agent to your metrics stack.
How do I add Metabase AI Assistant to my agent?
Register the metabase-ai-assistant npm MCP stdio server, set METABASE_URL and METABASE_API_KEY or username/password, and optional DATABASE_* vars for direct Postgres.