Fed’s FRED launches AI connector as data demand surges

Artificial intelligence is forcing the Federal Reserve’s FRED platform to rethink how economic data is accessed, interpreted and delivered, as AI agents become a major source of traffic, according to Fed Governor Christopher Waller.

At FRED Con 2026 on Oct. 1, Waller said the platform’s central value remains its role as a trusted conduit for data from governments and other sources, providing a consistent way to access and visualize information without commercial interests.

FRED, or Federal Reserve Economic Data, is a free online database and visualization platform operated by the Federal Reserve Bank of St. Louis. It offers hundreds of thousands of economic and financial time series covering areas such as inflation, GDP, employment, interest rates, the money supply and exchange rates.

FRED has evolved alongside technology for decades, from a mailed weekly data digest launched in 1961 to a computer bulletin board in 1991 and a website in 1995. The platform has since expanded from 865 data series to more than 850,000 and attracted 18 million unique users in 2025.

Waller said AI can make economic data more useful by allowing people to search, summarize and visualize information faster and at a much larger scale. It can also broaden access by explaining complicated economic concepts to people without specialized knowledge.

However, he said AI creates new risks because systems can generate content rather than simply distribute information. Those risks include hallucinated information, bias, hidden assumptions and inaccurate attribution, as well as analytical mistakes such as treating correlation as evidence of causation.

FRED is currently seeing traffic grow by 150% annually, with AI and bots responsible for most of the increase. About half of all FRED visits now come from AI agents, although the St. Louis Fed cannot always determine whether those agents are properly crediting the underlying data source or FRED.

FRED has launched an MCP Connector to make its data easier for AI systems to access and retrieve, allowing users to bring their own AI agent to the platform. Waller said the shift requires FRED to rethink how information is presented because AI agents consume data differently from human users and need information that can be accurately passed back to their users with appropriate documentation.

The St. Louis Fed is also developing capabilities specifically for AI agents and working to connect FRED with FRASER and ALFRED. FRASER can provide historical Federal Reserve documents, while ALFRED preserves data in earlier forms, including initial releases before revisions.

Waller said these efforts would support the next generation of data access while preserving FRED’s emphasis on reliability, transparency and trust.