Waller unveils FRED AI connector as bot visits reach 50 percent
Federal Reserve Governor Christopher J Waller announced the launch of the FRED MCP Connector to integrate economic data directly with artificial intelligence systems. Speaking at FRED Con 2026, Waller noted that automated AI agents now account for half of all platform traffic.
Automated bots claim half of platform visits
Traffic to the Federal Reserve Economic Data (FRED) platform is growing at 150 percent annually, with artificial intelligence agents and automated bots now generating roughly half of all visits.
Speaking at the St. Louis Fed, Federal Reserve Governor Christopher J Waller announced the rollout of the FRED MCP Connector, a tool using the Model Context Protocol that allows users to query over 850,000 data series through external AI models.
Waller also revealed that the Federal Reserve Board will soon route users seeking certain Board-published data directly to FRED.
The platform, which reached 18 million unique visitors in 2025, is expanding integration with the FRASER and ALFRED archival databases.
From dial-up modems to algorithmic queries
The database originated in 1961 as a typed weekly data digest distributed by St. Louis Fed Research Director Homer Jones, before shifting to telephone answering machines, 1991 dial-up bulletin boards, and a 1995 website with 865 data series.
While acknowledging that AI models accurately explain complex concepts, Waller warned of risks including data hallucination, missing source citations, and analytical errors such as confusing correlation with causation.
Embracing the algorithmic wave
Adapting FRED for machine consumption is a vital step as algorithmic data retrieval explodes.
Standardizing access via the Model Context Protocol secures attribution while streamlining workflows.
Yet central banks cannot prevent downstream AI models from misinterpreting core macroeconomic signals.