A faster, normalized alternative to the World Bank API
The World Bank, IMF, UN, FRED and OECD all publish their data for free, and it is the authoritative source. The pain is not the data, it is the integration: five different APIs, five auth schemes, SDMX vs nested JSON vs CSV, mismatched country and indicator codes, pagination quirks, and no uptime guarantee. Statistics of the World normalizes all of it into one schema: 335 indicators across 218 countries, as JSON or LLM-ready markdown, plus a hosted MCP server your agent can call directly.
How the options compare
| Provider | Sources | Auth | Agent-native (MCP) | Pricing | Best for |
|---|---|---|---|---|---|
| Statistics of the World | IMF + World Bank + UN + FRED + OECD (unified) | None under 1K/day, or a free key | Yes — hosted, 7 tools | Free; Pro $49/mo | One clean call across sources, or an AI agent |
| World Bank Indicators API | World Bank only | None | No (community single-source only) | Free | World Bank WDI series, if you parse the nested JSON |
| IMF Data (SDMX) | IMF only | None | No | Free | IMF/WEO macro series, if you can write SDMX |
| FRED | Federal Reserve (US-centric) | Free key (120 req/min) | No (official) | Free | US time series by series ID |
| DBnomics | Many (unified, academic) | None | No | Free | Researchers who know provider + dataset codes |
| Trading Economics | Broad + real-time markets | Sales-gated key | No | Enterprise (quote) | Funded teams needing real-time + SLA |
All sources above are reputable; the table compares integration effort and access pattern, not data quality.
Same task, less glue code
Get 20+ years of US GDP. With the World Bank API you look up the indicator code, handle the page-wrapped array-of-arrays response, and strip nulls. With SOTW it is one call that also returns YoY, CAGR, and min/max.
# Which indicator code is GDP again?
curl "https://api.worldbank.org/v2/country/US/\
indicator/NY.GDP.MKTP.CD?format=json&per_page=100"
# -> [ {pagination}, [ {year, value, ...}, ... ] ]
# then: drop page 0, filter nulls, sort, derive
# growth yourself.curl "https://statisticsoftheworld.com/api/v2/\
history?indicator=IMF.NGDPD&country=USA"
# -> { data:[{year,value,formatted}], yoy,
# cagr, min, max, coverageYears,
# relatedIndicators, suggestedQueries }
# add &format=markdown for an LLM-ready table.Use the World Bank API directly when
- You only need World Bank WDI series, nothing else.
- You do not mind the nested pagination and code lookups.
- You are doing a one-off pull, not shipping a product.
Use SOTW when
- You need IMF, World Bank, UN, FRED or OECD together, one schema.
- You are building an AI agent and want one MCP tool, not five gov APIs.
- You want rankings, comparisons, and history without writing the glue.
- You want a stable, versioned contract you ship against.
For AI agents: one MCP tool, not five gov APIs
The free government APIs have no first-class hosted MCP. SOTW is a hosted, multi-source MCP server, published to the official MCP registry, so an agent can compare any two countries on any indicator in a single tool call instead of learning SDMX. Add it to Claude Desktop, Cursor, or Cline:
{
"mcpServers": {
"sotw": {
"transport": {
"type": "streamable-http",
"url": "https://statisticsoftheworld.com/api/mcp",
"headers": { "X-API-Key": "sotw_YOUR_KEY_HERE" }
}
}
}
}The X-API-Key is optional under 1,000 requests/day. See /mcp for full setup.
Where the data comes from
SOTW does not generate data. It ingests, normalizes, and serves the primary sources, with source attribution and data vintage on every value:
Stop writing adapters for five APIs
Free up to 1,000 requests/day, no signup needed to try. A free key adds a usage dashboard; API Pro is $49/mo for 50,000/day plus Data Pro downloads. Provider reuse terms remain source-specific.