> ## Documentation Index
> Fetch the complete documentation index at: https://docs.agentkey.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Macroeconomic data for AI agents

> World Bank indicators, Eurostat datasets, US GDP, CPI, fed funds, Treasury yields, unemployment and payrolls, plus economic calendars for AI agents.

Macro data is public but awkward: every statistical office has its own codes,
dimensions and file formats, and an agent that wants GDP for three countries,
euro-area inflation and the latest payrolls number ends up learning three
query languages. Through AgentKey your agent asks for the series in plain
words — `find_tools`, `describe_tool`, `execute_tool` — with the World Bank,
Eurostat, AlphaVantage and Finnhub behind the scenes, and Tushare for China.

## What you can ask

* *"GDP per capita for Vietnam, Indonesia and the Philippines since 2000."*
* *"Compare euro-area and US inflation over the last five years."*
* *"What is the effective federal funds rate right now, and the 10-year Treasury yield?"*
* *"US nonfarm payrolls and the unemployment rate for the past three years."*
* *"Which upper-middle-income economies in Sub-Saharan Africa does the World Bank list?"*
* *"Germany's quarterly GDP from Eurostat, seasonally adjusted, in chained volumes."*
* *"What economic releases are due this week?"*
* *"中国最近两年的 CPI 和 PMI"*

## What's covered

**Global development indicators (World Bank)**

* Time series for GDP, GDP growth, GDP per capita, population, inflation,
  unemployment, exports and imports, FDI, government debt, exchange rates, life
  expectancy, poverty, Gini, internet users and thousands more, for one or
  several economies
* Indicator definitions with unit, source and methodology notes; indicator
  codes browsable by topic — Economy & Growth, Education, Energy & Mining,
  Environment, Health, Poverty and others
* Every economy and aggregate grouping with ISO codes, region, income level
  and lending type, plus single-economy profiles
* The list of underlying databases — World Development Indicators,
  International Debt Statistics, Global Economic Monitor and more

**European statistics (Eurostat)**

* Dataset cells filtered by country, period and dimension — GDP, inflation,
  unemployment, population, government debt, industrial production, trade,
  energy, emissions, poverty and thousands more — for EU member states, EFTA
  and candidate countries, returned as JSON-stat
* Dataset descriptions with full title, structure id and last-update
  timestamps
* Code lists for dimensions such as national-accounts item, unit, frequency,
  seasonal adjustment, sex and age

**United States**

* Real GDP (quarterly or annual) and real GDP per capita
* CPI (monthly or semiannual) and the annual inflation rate
* Effective federal funds rate; Treasury yields at daily, weekly or monthly
  frequency
* Unemployment rate, nonfarm payrolls, advance retail sales and durable goods
  orders

**Calendars and indicator feeds**

* Recent and upcoming economic events; economic indicators by code; the list
  of supported countries

**China**

* CPI, PPI, GDP, PMI, money supply, social financing, Shibor and LPR — see
  [China A-shares](/capabilities/finance/china-a-shares)

## How to use it

Pass the request in plain language and let the router pick the tool:

```text theme={null}
find_tools(q="GDP per capita for Vietnam since 2000")
find_tools(q="欧元区最近五年的通胀率")
find_tools(prefix="finance/worldbank")       # indicators, countries, topics, sources
find_tools(prefix="finance/eurostat")        # datasets, dataflows, code lists
find_tools(prefix="finance/alphavantage")    # US macro series
find_tools(prefix="finance/finnhub")         # economic calendar and indicator data
describe_tool("<returned Provider/Operation>")
execute_tool("<Provider/Operation>", { ...params })
```

Name the source when you have a preference ("World Bank", "Eurostat") and the
router ranks it first. Always `describe_tool` before executing — indicator and
dataset codes, dimension filters and per-call costs differ between providers.

> [!NOTE]
>
> **Building with code?** The same capability is available as a fixed-contract
> [Scenario Method](/api-reference/scenario/overview) with automatic provider
> failover:
> [`finance.macro.series`](/api-reference/scenario/methods/finance-macro-series).

## FAQ

> **How do I find the right indicator or dataset code?**
>
> For the World Bank, browse the indicator codes under a topic, then read the
> indicator's definition to confirm what it measures before pulling values.
> For Eurostat, read the dataset's description and the code lists for its
> dimensions to build the filter. The agent can do both steps itself from a
> plain-language request.

> **What frequencies and units come back?**
>
> Each series carries its own: US real GDP is quarterly or annual, CPI is
> monthly or semiannual, Treasury yields are daily, weekly or monthly. World
> Bank indicator metadata states the unit; Eurostat exposes unit and
> frequency as dimensions you filter on.

> **Which countries are covered?**
>
> The World Bank route covers every economy and aggregate it publishes;
> Eurostat covers EU member states, EFTA and candidate countries; the US
> series come from AlphaVantage; China's CPI, PPI, GDP, PMI, money supply and
> policy rates come from Tushare.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.