
Metabase Ai Assistant
A comprehensive bridge to Metabase that exposes 134 tools across database operations, dashboard management, and AI-powered SQL generation. You get direct access to the Metabase API for creating questions and dashboards, managing users and permissions, plus Postgres-specific operations like table creation, indexing, and VACUUM commands. The standout features are AI tools that convert natural language to SQL, optimize queries, and auto-generate dashboard templates. It includes read-only mode by default to prevent accidental destructive operations, activity logging for audit trails, and metadata analytics for tracking query performance and content usage. Reach for this when you need to automate BI workflows, generate SQL through conversation, or manage Metabase infrastructure programmatically without switching to the web interface.
- GitHub stars
- 43
- tools
- 8
- auth
- Secret needed
- transport
- STDIO
- env vars
- 9
- last updated
- Jan 23, 2026
Tools
8 toolsPublic tool metadata for what this MCP can expose to an agent.
dbt_inspect_modelsParses dbt manifest.json and MetricFlow semantic models.
Parses dbt manifest.json and MetricFlow semantic models.
No parameter schema in public metadata yet.
dbt_prioritize_sourcesDynamically routes natural language questions to pre-aggregated, tested dimensional and fact tables.
Dynamically routes natural language questions to pre-aggregated, tested dimensional and fact tables.
No parameter schema in public metadata yet.
semantic_memory_proposeProposes a business rule in PENDING_APPROVAL status.
Proposes a business rule in PENDING_APPROVAL status.
No parameter schema in public metadata yet.
semantic_memory_approveExplicitly activates the rule with required data steward comments.
Explicitly activates the rule with required data steward comments.
No parameter schema in public metadata yet.
semantic_memory_deprecateSafely soft-archives rules with mandatory audit reasons (DEPRECATED).
Safely soft-archives rules with mandatory audit reasons (DEPRECATED).
No parameter schema in public metadata yet.
semantic_memory_restoreInstantly restores archived rules.
Instantly restores archived rules.
No parameter schema in public metadata yet.
semantic_memory_listLists all rules with complete audit history and timestamps.
Lists all rules with complete audit history and timestamps.
No parameter schema in public metadata yet.
ai_sql_execute_and_healAutomated self-healing SQL execution with a 3-iteration error-recovery loop.
Automated self-healing SQL execution with a 3-iteration error-recovery loop.
No parameter schema in public metadata yet.
Sponsors
Configuration
METABASE_URL*Metabase instance URL (e.g., http://localhost:3000)
METABASE_USERNAMEMetabase username for authentication
METABASE_PASSWORDsecretMetabase password for authentication
METABASE_API_KEYsecretMetabase API key (alternative to username/password)
DATABASE_HOSTPostgreSQL database host for direct SQL execution
DATABASE_PORTPostgreSQL database port
DATABASE_NAMEPostgreSQL database name
DATABASE_USERPostgreSQL database user
DATABASE_PASSWORDsecretPostgreSQL database password
Readme
Metabase AI Assistant — Model Context Protocol (MCP) Server
Metabase AI Assistant is an enterprise-grade Model Context Protocol (MCP) server that connects Large Language Models (LLMs), AI coding assistants, and automated data workflows directly to your Metabase Business Intelligence instance.
Details
- Repository
- github.com/enessari/metabase-ai-assistant
- Category
- Databases
- Package
metabase-ai-assistant- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active