DataHub Announces Support for Model Context Protocol
See How Companies Like Block Enhance Data Discovery with AI Agents
AI agents are unlocking a powerful new way to explore and interact with data. By integrating generative AI applications into DataHub via the Model Context Protocol (MCP), teams can now query, understand, and act on data more naturally, right where they work.Ask questions about your data, investigate impact of potential changes, and identify stakeholders —all without leaving the tools you already use daily. Whether in Slack, your IDE, or dedicated AI interfaces, these agents make complex data discovery intuitive and efficient.
Watch the video below to see how Block is using AI agents with DataHub.
Benefits of Agentic AI Discovery
Democratize data with AI agents that can serve as intermediaries between users and complex data systems. Using AI agents lets you easily identify key stakeholders, understand the impact of changes with built-in analysis tools, and identify downstream effects of code changes with in-IDE visibility.
- Natural language Q&A - Ask questions about datasets, lineage, and metrics in natural language to quickly find the information you need, eliminating the need for specialized query languages or deep technical knowledge
- Metadata discovery and documentation assistance - Automatically identify, categorize, and document metadata, making institutional knowledge accessible to everyone
- Stakeholder identification and impact analysis - Use built-in analysis tools that map relationships and dependencies
- In-IDE visibility - See the downstream effects of code changes directly in your development environment
Access DataHub with AI Agents Anywhere You Work
This functionality is available across IDEs like Cursor, AI desktop interfaces such as Claude, and through Slack integration (available on DataHub Cloud only). The data team at Block is already leveraging this technology to drive data discovery, change management, and incident management.
AI Agents and MCP Roadmap
Open innovation is at the heart of this initiative, bringing together cutting-edge AI with open standards to make data more accessible, actionable, and impactful. At the forefront is DataHub’s implementation of the Model Context Protocol (MCP)– the open standard from Anthropic that is designed to make tools discoverable and queryable by AI agents in a consistent and scalable way. Building on top of that is the open source AI agent from Block called Goose, that enables users to connect large language models (LLMs) to real-world actions.
We’re also investing in the building blocks that underlie the DataHub MCP server. For example, we’ve intentionally designed our SDKs to be easy to use for humans and AI agents alike. We’re also working on expanding the authentication options available with the DataHub MCP server.
We're excited about the improved data democratization this will bring by turning passive metadata into an active resource, empowering both data engineers and business users to interact with complex data systems through intuitive, AI-powered natural language interfaces.
Get a Demo of DataHub Cloud
Experience the premium version of DataHub with observability and governance built-in. Sign up to see DataHub Cloud's unified approach to Data Discovery, Governance, and Observability in action.
