Install guide and config at curatedmcp.com Databricks MCP: Give Claude Direct Access to...

Install guide and config at curatedmcp.com

Databricks MCP: Give Claude Direct Access to Your Lakehouse

The official Databricks MCP server bridges the gap between AI agents and your data lakehouse. Instead of describing your schema or manually copying query results, Claude can now execute notebooks, run SQL directly against Delta Lake, manage clusters, and inspect data lineage—all conversationally.

This is particularly useful if you're building AI workflows that need to touch production analytics or data engineering assets. Your agent can query tables, check job status, and pull insights without leaving the conversation.

What It Does

Databricks MCP unlocks several concrete capabilities for AI agents:

• Execute notebooks and fetch results — tell Claude to run a specific notebook, and it returns the output directly
• Query Delta Lake and Unity Catalog — ask questions about your lakehouse data; Claude runs the SQL
• Manage compute — inspect cluster status, scale resources, or review job history
• Inspect schemas and lineage — Claude can explore table structure and data provenance
• Access MLflow — query model registry, review experiment runs, and log metrics programmatically
• Generate analytics — run SQL, build insights, and create simple visualizations

The server handles authentication via your Databricks workspace token, so no credential juggling required once configured.

How to Install

Install the server via npm:

npx -y @databricks/mcp

Add it to your Claude Desktop config:

{
  "mcpServers": {
    "databricks-mcp": {
      "command": "npx -y @databricks/mcp",
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.cloud.databricks.com",
        "DATABRICKS_TOKEN": "your-personal-access-token"
      }
    }
  }
}

See the official docs for detailed setup across all platforms.

Real-World Use Cases

• Automated data quality checks: Ask Claude to run your validation notebook nightly, then alert you if row counts or schema mismatches appear
• Self-service analytics: "Query revenue by region from the silver tables and explain trends" — Claude executes the SQL and interprets results
• ML experiment tracking: Have Claude log hyperparameters and metrics to MLflow during training, or query past experiments to compare models

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.