FRED MCP Connector: Connect AI Assistants to FRED Data

FRED MCP Connector: Connect AI Assistants to FRED Data

FRED MCP Connector

Bring FRED into the conversation with your favorite AI applications and tools.

What is the FRED MCP Connector?

The FRED MCP Connector is the official Model Context Protocol (MCP) server for FRED, the economic database maintained by the Research Division of the Federal Reserve Bank of St. Louis. It lets AI assistants such as Claude search FRED, retrieve series data and metadata, create FRED graphs and generate CSV or JSON downloads. The connector is free, uses a FRED MCP Connector account and does not require a FRED API key.

Questions? Contact us.

  • Provider: Federal Reserve Bank of St. Louis (official)
  • MCP endpoint: https://mcp.stlouisfed.org
  • Cost: Free
  • Account: FRED MCP Connector account (separate from a FRED user account)
  • FRED API key: Not required
  • Connector directories: Claude
  • Tools: 5 (search, metadata, observations, download link, graph)

Official connector vs. community servers

Is this the same as other FRED MCP servers?

No. The FRED MCP Connector is the only MCP server operated by the Federal Reserve Bank of St. Louis. Several independent, open-source projects also connect AI tools to FRED; they’re maintained by third parties, typically require you to install software and supply your own FRED API key, and aren’t affiliated with or supported by the St. Louis Fed. To use the official connector, add https://mcp.stlouisfed.org or install FRED from your AI assistant’s connector directory.

How do I connect FRED to my AI assistant?

MCP endpoint (you do not need to navigate to this URL in your browser): https://mcp.stlouisfed.org

If FRED appears in your AI assistant's connector directory, you can add it from there. 

Directory links currently available are here:

Want to be notified when it is available in your AI assistant’s connector directory? Sign up here.

Otherwise, add it as a custom connector using the MCP endpoint URL above. Instructions for adding custom connectors are available at your AI provider of choice.

When you connect, you'll be redirected to sign in.

  • New MCP Connector users: Create a free FRED MCP Connector account by clicking the "Sign up" link at the bottom of the Okta view below. You'll be asked to accept the Terms of Service and Privacy Policy once, at signup.
    • Already have a FRED account? The FRED MCP Connector uses its own account. Your existing FRED account (the one used for an API key, saved graphs or dashboards) won’t sign you in here, so choose “Sign up” the first time you connect.
  • Returning MCP Connector users: Sign in with your existing account. You won't be asked to accept the terms again.

Okta prompt for the FRED MCP connector

Authorize your AI assistant to access FRED on your behalf, and you'll be returned to the AI tool text prompt.

You only create one account. If you use FRED from more than one AI assistant, sign in with the same account in each. Each assistant needs its own authorization, but not a separate signup.

No FRED API key is required.

Getting started

After connecting FRED to your AI assistant, try asking:

What is the latest unemployment rate?

or:

How has inflation changed over the past five years?

Your assistant uses the FRED tools to find appropriate economic series, retrieve their data, and answer your question.

What tools does the FRED MCP Connector provide?

The FRED MCP Connector provides five tools:

  • Search series (fred-search-series): Search FRED for economic time series using keywords or concepts. Example: “Find a FRED series for prime-age labor-force participation.”
  • Series metadata (fred-get-series-metadata): Units, frequency, seasonal adjustment, source and notes. Example: “What exactly does the UNRATE series measure?”
  • Series observations (fred-get-series-observations): Data for a series, with optional date range, transformation or frequency aggregation. Example: “Give me monthly unemployment since 2020.”
  • Download link (fred-download-series-observations-link): A downloadable CSV or JSON file with the same options. Example: “Download year-over-year CPI inflation since 1990 as CSV.”
  • Graph series (fred-graph-series): An interactive FRED graph with up to five series. Example: “Graph unemployment and labor-force participation since 2000.”

Where does the data come from?

FRED is maintained by the Research Division of the Federal Reserve Bank of St. Louis. It brings together economic time series from Federal Reserve entities, U.S. government statistical agencies (such as the Bureau of Labor Statistics, Bureau of Economic Analysis, and Census Bureau), international organizations, and other providers. Series metadata identifies the underlying source where available.

Do I need to know series IDs?

Every FRED series has an identifier, such as:

  • UNRATE - Unemployment Rate
  • CPIAUCSL - Consumer Price Index for All Urban Consumers
  • FEDFUNDS - Federal Funds Effective Rate

You can use these when you know them, but you usually don't need to. Instead of "Get UNRATE," just ask "What is the U.S. unemployment rate?" Your assistant searches for an appropriate series and retrieves it.

Search results include the series ID, units, frequency, seasonal adjustment, source, release, historical coverage, and notes - which helps the assistant choose among similar measures.

Getting and analyzing data with the FRED MCP Connector

Ask directly for current or historical data:

What is the latest unemployment rate? Show unemployment since 2000. How has labor-force participation changed since 1990?

FRED can also transform data during retrieval:

What is the year-over-year CPI inflation rate? How much did payroll employment change each month this year? What was real GDP growth?

The right calculation depends on how a series is measured, and your assistant inspects the series metadata before choosing one. For example, an unemployment rate expressed in percent is already a rate, while a price index is not itself an inflation rate - inflation is calculated from changes in the index.

Comparing economic indicators

Compare CPI and PCE inflation. Compare unemployment with job openings. How have wages behaved relative to inflation?

The assistant retrieves the relevant series and analyzes them together. When similar measures differ in definition, it can use FRED metadata to explain the differences.

Creating graphs

Graph the unemployment rate since 2000. Plot mortgage rates and the federal funds rate together.

FRED graphs can contain up to five series, transformed individually and displayed on left or right axes.

Downloading data

Download unemployment since 1980 as CSV. Give me CPI inflation since 1990 as a CSV.

FRED generates downloadable CSV or JSON files for use in spreadsheets, statistical packages, and programming environments.

Changing data frequency

Some series are daily or weekly. FRED can aggregate them to a lower frequency:

Give me monthly averages of this daily interest-rate series.

Aggregation methods are average, sum, and end of period; the right choice depends on what the series measures.

Reading the data

A few FRED conventions worth knowing:

  • A “.” in the observations means the value is missing for that period - it is not zero.
  • The date on an observation is the period it measures, not the date it was published or retrieved.
  • Responses include realtime_start and realtime_end, which record the as-of date of the data. Economic statistics are revised, so the same observation can change between retrievals.

Example prompts

  • What is the latest unemployment rate?
  • What do recent data say about the U.S. labor market?
  • How has core inflation changed recently?
  • Compare CPI and PCE inflation.
  • Graph the federal funds rate since 1990.
  • What happened to unemployment in previous recessions?
  • Find a FRED series for house prices.
  • Download monthly unemployment since 1980 as CSV.

You can also follow up naturally:

How has inflation changed recently? What about core inflation? Graph both since 2019.

Your assistant combines the FRED tools as needed to complete the analysis.

Frequently asked questions (FAQ)

Is the FRED MCP Connector free?

Yes. The connector and the FRED data it returns are free to use; you need a free FRED MCP Connector account.

Do I need a FRED API key?

No. The connector signs you in with your FRED MCP Connector account instead.

Which AI assistants can use the FRED MCP Connector?

Any assistant that supports custom MCP connectors can connect using https://mcp.stlouisfed.org. FRED is listed in the Claude connector directory today: https://claude.ai/directory/fred.

We will add other directory links on this page as they go live.

Why doesn’t my existing FRED account work?

The connector uses a separate account. Choose “Sign up” the first time you connect, then use that account in every assistant.

Why did a value change between two answers?

Many economic statistics are revised after release. Each response records an as-of date (realtime_start and realtime_end). For earlier vintages of a series, use ALFRED.

Does the FRED MCP Connector provide financial advice?

No. It provides access to economic data. It does not provide financial, investment, legal or policy advice.

How do I report a problem or ask a question?

Contact the FRED team.

Revisions and limitations

Many economic statistics are revised after initial publication, so values retrieved today may differ from earlier retrievals. When timing matters, consider the observation date, series definition, and underlying source.

AI-generated analysis can contain mistakes. For important research or decisions, verify the relevant series, units, dates, and definitions.

FRED MCP provides access to economic data. It does not provide financial, investment, legal, or policy advice.

Sources and provenance

Maintained by: FRED team, Research Division, Federal Reserve Bank of St. Louis.
Data sources: Series metadata identifies the original source (for example, BLS, BEA, Census).
Citing data retrieved through the connector: see Cite the Data.
Questions: Contact us.