Waller links FRED to AI agents as bot traffic grows 150 percent
Federal Reserve Governor Christopher Waller announced the launch of the FRED MCP Connector to link central bank data directly to artificial intelligence systems. Speaking in St. Louis, Waller noted that automated bots now account for half of all visits to the database.
Adapting to machine visitors
Federal Reserve Governor Christopher Waller presented the FRED MCP Connector at the St. Louis Fed, establishing a standardized interface using the Model Context Protocol.
The tool allows external artificial intelligence agents to retrieve structured economic data directly.
Annual traffic to FRED is currently expanding by 150 percent, with automated bots already generating approximately 50 percent of all platform visits.
Waller also stated that users seeking specific data published by the Federal Reserve Board will soon be rerouted to FRED.
The platform currently hosts more than 850,000 economic data series from global sources and reached 18 million unique users in 2025.
From paper mailings to archival links
FRED originated in 1961 as a mailed paper digest created by St. Louis Fed Research Director Homer Jones before launching online in 1995 with 865 data series.
While AI tools accelerate data summarization, Waller highlighted operational risks, including algorithmic hallucinations and flawed causal interpretations.
To enhance machine retrieval, the St. Louis Fed is connecting FRED with the FRASER historical archive and the ALFRED vintage database.
“FRED is the goodwill ambassador for the Federal Reserve,” Waller said.
A necessary pivot to machine consumption
Building direct pipelines for AI models acknowledges the reality that bots now dominate data consumption.
The initiative preserves the Fed's authority as a primary source while minimizing the risk of automated misattribution.
Technical integration cannot eliminate algorithmic hallucinations, but structured access ensures foundational data integrity.