St. Louis Fed Adapts FRED Database for AI Agents

Federal Reserve Economic Data (FRED) has unveiled a FRED MCP Connector that employs the Model Context Protocol (MCP) to streamline integration between FRED data and artificial intelligence systems, announced Federal Reserve Board Governor Christopher J. Waller on October 1 during FRED Con at the Federal Reserve Bank of St. Louis.

Waller explained during his speech that this new connector allows users to employ their preferred AI agent for retrieving and accessing data from FRED. Waller emphasized that this development is significant as FRED, traditionally designed for human consumption, now needs adaptation to cater to the unique way AI bots perceive and interpret data.

He stated that the FRED team must modify how FRED presents data to ensure accurate and well-documented transmission to the AI agents' masters. FRED, a comprehensive online database housing hundreds of thousands of economic data time series from diverse sources, is developed and maintained by the St. Louis Fed’s Research Department.

The Model Context Protocol, an open standard introduced by Anthropic in 2024, aims to enhance AI systems' utility in real-world business scenarios by enabling their connection to various business tools, databases, and workflows, according to PYMNTS in November 2025. Waller cautioned that AI's potential inaccuracies in citing data sources and misinterpretation of data could challenge FRED's reputation as a dependable data source created by external organizations.

Currently, half of FRED's visits are from AI agents seeking data. Despite generally accurate performance, AI agents occasionally fail to attribute data sources correctly. Waller assured that the FRED MCP Connector will ensure AI bots accurately convey retrieved data. He also mentioned that FRED staff is developing new skills to help AI agents better understand and work with economic data, ensuring improved accuracy and effectiveness.