---
title: "Databricks MCP: named agents, logged calls | Rig"
description: "Databricks MCP for Claude, Cursor and your own agents, with a role, token and daily call limit per agent. Rig Gateway logs every query under the agent’s name."
canonical: "https://rig.so/mcp/databricks"
format: markdown
---

MCP

# I’d want every Databricks MCP call named and logged

Databricks MCP connects AI tools to Genie, Databricks SQL and Unity Catalog through the Model Context Protocol. Rig adds Gateway on top, so each agent that reaches Databricks has a role of its own, a token, a daily call limit and an audit trail under its name.

By [Rig](https://rig.so/about) · Last updated 20 November 2026

[Book a call with an architect](https://rig.so/book-demo)

## At a glance

- Databricks runs managed MCP servers for Genie, Databricks SQL, AI Search and Unity Catalog functions, plus Unity Gateway MCPs governed with grants and policies.
- Status varies by server, from generally available to beta and public preview, so check the Databricks page for the one you plan to use.
- Rig connects to a Databricks SQL warehouse read-only, carries Unity Catalog grants through, and adds a role, token and call cap per agent.

## What is Databricks MCP?

Databricks documents two families of managed MCP servers. Unity Gateway MCPs live in the system.ai schema, including Genie One and Databricks SQL, and are governed with grants and policies. Workspace endpoints cover a Genie Agent, AI Search, a legacy Databricks SQL endpoint and Unity Catalog functions, each at its own URL, and Databricks notes that calls to these bypass Unity Gateway. Genie One MCP and Unity Gateway are generally available. Some system.ai servers are in beta, and the workspace endpoints are in public preview.

## How do Databricks managed MCP servers handle access?

Each server uses OAuth, with a scope per server family. The Databricks SQL server runs with the caller’s Unity Catalog and warehouse permissions, and the workspace endpoints rely on permissions on the underlying resource. For a person signing in, that is the right model. An agent that a team shares, or one that runs on a schedule, still needs an identity that belongs to the agent, so its access can be narrowed and its calls traced back to it.

## Why do agents on Databricks need their own role?

Unity Catalog can tell you which principal ran a query. If every agent runs as the same service principal, or as the engineer who set it up, the log cannot tell a fraud agent from a forecasting agent. A role per agent keeps each one to the catalogs and columns its job needs, and a daily cap protects the SQL warehouse from an agent that keeps retrying.

## How Rig does it on Databricks

Rig connects to a Databricks SQL warehouse with a read-only identity, reads Unity Catalog, and builds Rig Map over your tables, joins and metric definitions. Unity Catalog grants and column masking carry through, and Gateway adds its own layer: a role per person and per agent, schema and column rules, PII masking, a daily call cap and a named audit trail. Queries run as Databricks SQL on your own compute. Before an agent writes SQL it looks up certified metrics in Rig Map, and Rig validates the query before it runs.

## What about Unity Catalog metric views and Genie?

Metric views are Databricks’ semantic layer, defined in YAML and usable from SQL, dashboards and Genie. Genie answers inside the Databricks workspace. If your team also asks in Claude or Cursor, Rig Map gives those tools the same definitions, and Gateway applies one set of rules across every tool connected through Rig.

## Who runs governed agents on Rig?

Novakid’s sales managers and sales ops investigate their own numbers in Claude through Rig, on a warehouse of more than 5,000 tables narrowed to tables the team trusts. The team has made 22,412 requests since March. Novakid runs BigQuery. The same Gateway model applies on Databricks.

## Common questions

What is Databricks MCP?

It is Databricks’ set of managed MCP servers that expose Genie, Databricks SQL, AI Search and Unity Catalog functions to MCP clients, governed by Unity Catalog permissions and, for Unity Gateway MCPs, by grants and policies.

How do I give agents Databricks access with a named role and a log?

Create the agent in Rig from a role you trust, narrow its catalogs, columns and masked fields, set its daily call cap and paste its MCP URL and token into the tool. Every call lands in Rig’s audit log under the agent’s name.

Do Unity Catalog permissions still apply through Rig?

Yes. Rig connects read-only, Unity Catalog grants and column masking carry through, and Rig’s own role, column and row rules apply on top.

How do I expose governed business metrics to Claude and Cursor?

Certify the metrics once in Rig Map, then connect Claude or Cursor to Rig’s MCP server. A person signs in with OAuth and gets their own role. An agent gets its own role and token. Either way, the tool looks up the certified metric before it writes SQL, sees only the schemas, columns and rows the role allows, and every call lands in the audit log.

## Related guides and services

[data context and governance](https://rig.so/use-cases/data-context-and-governance)[gateway](https://rig.so/products/gateway)[rig map](https://rig.so/products/rig-map)[mcp](https://rig.so/uses/mcp)[databricks context layer and metric views](https://rig.so/blog/databricks-context-layer-and-metric-views)[how to connect ai agents to your data warehouse](https://rig.so/blog/how-to-connect-ai-agents-to-your-data-warehouse)[databricks](https://rig.so/integrations/databricks)[novakid](https://rig.so/case-studies/novakid)

## Sources and product details

Vendor information checked on 20 November 2026. Confirm current editions, licence terms and prices with the vendor before buying.

- [Databricks: managed MCP servers](https://docs.databricks.com/aws/en/generative-ai/mcp/managed-mcp)
- [Databricks: Unity Catalog metric views](https://docs.databricks.com/aws/en/uc-semantics/metric-views/)
- [Model Context Protocol specification](https://modelcontextprotocol.io/specification)
- [Rig guide: give your AI agents their own access](https://rig.so/guides/agent-roles)

Talk to Rig

## What work do you want to improve?

Tell us where your team spends time, loses money or needs better information.
In a 30-minute call, we’ll discuss the data and systems work that could help.

[Book a call with an architect](https://rig.so/book-demo)Use the platform with your team, or add our engineers.

---

- [This page as HTML](https://rig.so/mcp/databricks)
- [Site map for language models](https://rig.so/llms.txt)
- [API and agent documentation](https://rig.so/developers)
