Treasure Data, a provider of enterprise-scale Customer Data Platforms (CDP) powered by AI, has announced the release of its new open-source MCP Server. This connector enables AI assistants such as Claude, GitHub Copilot Chat, and Windsurf to interact directly with a Treasure Data environment, transforming the CDP into a conversational interface.
The MCP Server, built on the open Model Context Protocol (MCP), provides data teams with enhanced capabilities for exploring and analyzing customer data. Users can now leverage plain language within a conversation window to query segments, explore tables, and conduct data analysis, thereby making data insights more accessible. This development aims to overcome traditional bottlenecks in data access, such as reliance on complex query languages, disconnected tools, and manual workflows.
Rafa Flores, Chief Product Officer of Treasure Data, stated, “This is the beginning of natural language analytics at scale. We’re making it radically easier for teams to get value from their data without needing to write complex SQL or rely on someone else to do it for them. This is a major shift in how teams interact with data: Just you, your questions, and your AI.”
The MCP Server functions as a local bridge between LLM-enabled tools and the Treasure Data platform. Once configured, it allows AI agents to securely interact with the CDP through structured tool calls. For instance, a user in Claude Desktop or GitHub Copilot in Visual Studio Code could input: “@treasuredata Find the top 5 most viewed products in the past 7 days and show the revenue by region.” The assistant then connects via the MCP Server to Treasure Data, generates the appropriate SQL, executes it, and returns results within the chat interface, eliminating the need for tab switching, manual query building, or SQL knowledge. The system also supports operations like describing table schemas, switching databases, and performing advanced analyses, such as identifying IP addresses with high unique page access within a specific timeframe. The MCP Server manages permissions, safely limits results, and securely handles API keys and environment variables.
While the MCP Server is read-only by default for secure data exploration, advanced users can enable write access for interactive workflows involving data inserts and updates. Its accessibility extends across various teams:
* **Marketing analysts** can swiftly assess campaign performance by region.
* **Data engineers** can validate new tables or perform point-in-time quality assurance (QA) checks.
* **Product managers** can explore user journey data and conversion funnels without requiring support from a BI team.
* **AI developers** can chain multi-step queries to train models and test hypotheses using natural language, without needing SQL fluency.
This innovation addresses a significant challenge for enterprises: the effective use of AI is often hindered by the inability of Large Language Models (LLMs) to access high-quality, governed data. The Treasure Data MCP Server removes this barrier by enabling direct, secure, and intelligent access for AI to the CDP, facilitating more productive interactions with customer data.
The MCP Server is currently available in public preview on npm and is open for contributions and feedback on GitHub.
Treasure Data, based in Mountain View, California, operates as an Intelligent Customer Data Platform (CDP) designed for enterprise scale and powered by AI. The company empowers organizations to deliver hyper-personalized customer experiences at scale, aiming to increase revenue, reduce costs, and build trust.
