Looker debuts MCP Server to broaden AI developer access to data

As companies integrate AI into their workflows, connecting new tools to their existing data while ensuring consistent security and accuracy becomes increasingly important. We’re introducing Looker Model Context Protocol (MCP) Server, an integration in the MCP Toolbox for Databases. This allows AI applications such as chatbots and custom agents to connect to trusted data from the environments AI developers use every day.

Looker already helps thousands of organizations to access, analyze, and act on a single, consistent, and governed view of their data through its robust semantic layer, connecting to hundreds of data sources such as BigQuery, AlloyDB, and Cloud SQL. With the launch of the Looker in MCP Toolbox, we are extending our leadership in trusted generative AI for BI by bringing this functionality to the emerging world of AI applications and agents.

MCP is an open standard technology that allows large language models (LLMs) and AI applications to access other products consistently and securely. Looker’s MCP Toolbox integration connects applications to the LLM along with structured metadata and specific request parameters. In addition, the MCP Server can expose unstructured natural language information about how the data source is called and what type of information it returns.

MCP essentially acts as a universal translator, enabling AI models to:

• Discover and use tools dynamically: Rather than hardcoded integrations, AI agents can identify and interact with available capabilities in real-time.
• Access relevant, up-to-date context: AI models can pull live, verified information directly from its source, significantly reducing hallucinations and improving response accuracy.
• Ensure secure and governed data access: MCP provides a host-mediated security model, allowing fine-grained control over what data AI agents can access and how.

Intelligent AI apps, meet intelligent data

The debut of Looker’s MCP Server, combined with its semantic layer, transforms the opportunity for data-driven AI. There is no need for AI to write SQL. The AI queries Looker’s semantic layer and Looker generates the correct, optimized SQL. Here’s what this means for your organization:

1. Trusted data for AI, on-demand: Looker’s semantic layer ensures that all your business metrics and definitions are consistent and governed. With the Looker MCP Server, AI agents can now directly query this single source of truth, receiving accurate and reliable data-driven insights without the risk of misinterpretation or outdated information.
2. Enhanced security and data governance for AI: Looker’s MCP Toolbox integration inherits Looker’s robust security model, allowing administrators to define precise access controls for AI agents. You can dictate which AI applications can access what data, at what granularity, and for what purpose, all within the familiar Looker environment. Sensitive data remains protected, and audit trails ensure compliance.
3. Accelerated AI application development: Developers building AI-powered applications that need to interact with enterprise data often face complex integration challenges. By exposing Looker’s rich data models via a standardized MCP interface, AI developers can now easily connect their agents to a pre-defined, trusted data layer, reducing development time and effort.
4. Integration with the tools your AI developers use: Looker’s MCP Toolbox integration can be accessed today through any agent that supports MCP, including offerings such as Gemini’s Command Line Interface, Anthropic’s Claude Desktop, and Cursor.

Get started with Looker MCP Server

Looker’s MCP Server via MCP Toolbox continues Google Cloud’s commitment to making data truly useful and accessible, for all users, including the next generation of intelligent AI applications. We believe this release will empower organizations to unlock unprecedented value from their data through modern AI tools, driving smarter decisions and accelerating innovation across every facet of their business.

To get started with Looker MCP Server, check out our Quickstart guide on Github.