Airbyte, creator of the open data movement platform, today announced new capabilities that enable users to chat directly with their data, and use agents and AI assistants to monitor Airbyte pipelines, diagnose and resolve issues.

The new Model Context Protocol (MCP) server added to Airbyte’s data movement platform makes all of this possible. Using the MCP open source standard provides a natural-language interface to use with AI assistants such as Claude, Cursor, and ChatGPT.

“MCP technology enables all sorts of possibilities for data engineers, who can now simply direct questions about syncs and data sources to AI agents,” said Michel Tricot, co-founder and CEO of Airbyte. “The result is a superior experience and ability to do more faster. This fulfills our users’ requests for MCP, which is the most requested feature we’ve seen in our company’s more than 5-years history.”

For users of Airbyte Cloud and Airbyte Enterprise Flex, the MCP server also provides access to the new Airbyte context layer. The context layer supports direct queries on Snowflake and BigQuery destinations, with data synced from more than 50 Airbyte source connectors. Users can query their data warehouse in natural language from Claude or ChatGPT asking, for example, ‘what customer accounts have open support tickets?’

Additionally, Airbyte users managing data pipelines are able to ask when a dataset last synced successfully, identify and fix broken syncs, create and modify data connections, and more. For example, data engineers can ask Claude, ‘Why is my sync failing?’ and get a specific cause and recommended fix so that they don’t have to manually go through logs to find answers.

The open source Airbyte MCP server is available across Airbyte Cloud, Airbyte Enterprise Flex and self-managed Airbyte deployments. The MCP server is simple to install and authenticates with existing user credentials. The MCP server works with MCP-compatible clients, including ChatGPT, Claude, VS Code and other development environments and AI agents that support the protocol.

Also, the MCP Server governs agents to prevent rogue behavior. For example, deleting data connection sources or destinations requires confirmation, the MCP server never elevates user access privileges, and an audit log identifies all MCP originated actions.

Extending Airbyte for agentic data infrastructure

The MCP server establishes an agent-accessible operational layer for Airbyte. Initial capabilities focus on connecting agents securely, troubleshooting data syncs, monitoring connection health, and performing governed corrective actions. Airbyte plans to expand capabilities to support conversational management of sources, destinations, connections, jobs, users, workspaces, and other Airbyte resources.

The Airbyte MCP Server complements Airbyte’s APIs, user interface and infrastructure-as-code workflows. It is intended to give teams another secure way to operate Airbyte with the AI tools that they use increasingly, rather than replacing established management interfaces.

Availability

The Airbyte MCP Server is available now. Go here to learn more. For setup instructions, supported MCP clients, and product documentation, read the MCP docs.

About Airbyte

Airbyte is the open standard in data connectivity for AI and analytics. Its open-source engine, 700-plus connectors, and Connector Builder move data from any source into data warehouses or data lakes, with a context layer for AI agents. More than 25,000 companies, including 20% of the Fortune 500, run 2 million data pipeline syncs a day on Airbyte. Airbyte is available as a managed cloud instance, hybrid deployment model, or self-hosted open-source platform. Airbyte has been open source since 2020 and supports SOC 2 Type II, GDPR, and HIPAA Conduit compliance.

Media gallery

About The Author
Business Wire