Trusted agents at scale: Collibra launches MCP server in the Databricks Marketplace to power governed AI

Share on:

Databricks recently announced a significant leap forward in AI development with the launch of Agent Bricks, the Model Context Protocol (MCP) Catalog and the Databricks Marketplace. This new suite of tools empowers organizations to build and deploy sophisticated AI agents that can reason over data and connect to external services to take real-world actions. We are proud to be a partner in the Databricks Marketplace, bringing the Collibra Platform into this new AI ecosystem.

This partnership helps organizations solve a critical challenge in the new agentic AI era: how to ensure that AI agents operate with trusted, governed and well-understood data and tools.

Why Databricks and Collibra for the agentic AI era

As enterprises race to build AI agents, the ability for these agents to discover and use a vast array of internal and external tools is paramount. However, without a robust governance framework, this power creates significant risk. Without well-understood and high-quality data, the impact of AI initiatives is reduced. How can you ensure an agent uses the "certified" customer API versus an outdated, unmanaged one? How does an agent understand the context, quality and policies associated with data before acting on it?

This is where the joint value of Collibra and Databricks shines. Databricks provides the platform to build, manage and deploy AI agents and connect them to a universe of tools via MCP. Collibra provides the essential layer of context and governance about those tools and the data they access. This powerful combination enables organizations to accelerate AI development while maintaining the trust, compliance and security their business demands.

Discover and govern agent tools with the Collibra MCP server

Available now in the Databricks Marketplace, the new Collibra MCP server acts as a direct line of communication between your AI agents and the Collibra Platform.

When an AI agent built with Databricks Agent Bricks needs to find a tool or data asset, it can now query the Collibra MCP server to enrich its understanding. This allows the agent to get real-time answers to critical governance questions before it acts:

What AI use-cases are using this data set?
Is this data product certified? What is its data quality score?
Who is the owner or steward I can contact for more information?
Does this tool access sensitive data (e.g., PII)? What are the usage policies?

By integrating Collibra's rich metadata, lineage and governance context directly into the agent's operational loop, we move beyond simple tool discovery. We enable governed discovery, ensuring agents make smarter, safer and more compliant decisions.

Building trust directly into Databricks Agent Bricks

The integration of Collibra with Agent Bricks transforms how enterprises can safely deploy agentic workflows. Imagine a marketing agent tasked with identifying high-value customers in the EMEA region for a new campaign.

Using the Databricks MCP Catalog, the agent discovers several available APIs and data services related to customer information.
Before acting, the agent calls the Collibra MCP server to gather intelligence on these services.
Collibra instantly provides the context: one API is "Certified Gold," refreshed daily and approved for marketing use. Another is a "deprecated" version, and a third is a "development sandbox" with poor data quality.
Armed with this governance context, the agent confidently selects the certified, high-quality API to complete its task accurately and in compliance with company policy.

This seamless, machine-speed governance check prevents agents from using unreliable data, reduces the risk of hallucinations based on poor inputs and provides an auditable trail of how and why an agent made a particular choice.

Transform your data into a strategic AI asset

The data management landscape is evolving faster than ever. To maintain a competitive edge, organizations must not only manage their data but activate it intelligently. The combination of Collibra and Databricks provides a clear path to do just that.

With the combined power of Databricks' agentic framework and Collibra's native integrations for cataloging, governance, quality, and lineage providing the critical context agents need to act reliably, you can accelerate the process of building enterprise-grade AI. This frees your data and AI teams from the manual, repetitive tasks of validating tools and data sources across platforms, enabling them to spend more time innovating and creating value. This newfound freedom is when the real transformation of data into a strategic AI asset begins.

Learn more about the Collibra and Databricks partnership on our Databricks Partner Page.

You can find the Collibra MCP server in the Databricks Marketplace today. Stay tuned as we will continue to add tools to our offering, increasing the breadth of capabilities of the Collibra MCP server.

Larry Hau

Larry Hau

Vice President, Product Management

Collibra
Kiranmai Thotakura

Kiranmai Thotakura

Sr. Partner Engineer, Product

Databricks