StatCite — Economic Data & Stat Verification (with citations)

Official economic statistics where every number ships with its citation — plus a verifier that checks any claimed figure against the official series. Built and curated by a professional economist.

AI agents and analysts constantly need macro numbers they can defend: GDP growth, inflation, unemployment, government debt, exchange rates. This actor returns the official value and the receipt — source, dataset, series ID, canonical URL, license, retrieval date, and a ready-to-paste citation sentence.

What it does

Operation	What you get
indicator	Values for any of 42 active curated indicators (inflation, GDP growth, debt/GDP, current account, FDI, …) for 200+ economies, with citation
verify	Check a claimed figure against the official series → match / close / mismatch / cannot_verify, with diagnostics (wrong year? percent-vs-decimal? millions-vs-billions?) and the correct citable value
snapshot	11 headline indicators for one country, each with its own citation — instant country brief
inflation	"What is $100 (1995) worth in 2025 money?" — CPI-ratio method, formula disclosed, any country
fx	Currency conversion: ECB daily reference rates (~30 majors since 1999) + official annual rates for ~90 more currencies (BBD, XCD, JMD, KES…)
series	Raw access: worldbank/CODE, dbnomics/IMF/WEO:latest/...
search / registry	Discover indicators and datasets

Sources

World Bank World Development Indicators (CC BY 4.0) · IMF World Economic Outlook & Fiscal Monitor via the IMF's own DataMapper API (current vintage, DBnomics as fallback) · BIS central bank policy rates · ECB euro reference rates and euro-area HICP. Data is fetched live from official APIs — this actor adds no numbers of its own, and labels IMF projections as projections.

Source data is free at the source

The underlying data is available free of charge from its publishers — the IMF, the World Bank, the BIS and the ECB all publish it at no cost, and StatCite's own API at statcite.com is free and needs no key. What this actor charges for is the convenience layer: a managed, rate-limited, pay-per-event endpoint with the citation, verification and diagnostic work already done, billed per successful item.

This notice is also a licence obligation: the IMF's terms for published statistical data require that where the data is sold as part of a product, purchasers are told it is available free of charge from the IMF. Every response additionally carries the source's own licence and required attribution string in its citation object, so those terms travel with the data.

Example: verify a stat before publishing

Input:

{ "operation": "verify", "indicator": "inflation_cpi", "country": "USA", "period": "2023", "claimed_value": 4.1 }

Output (dataset item):

{
  "ok": true,
  "result": {
    "verdict": "match",
    "official_value": 4.116338,
    "explanation": "Claimed 4.1 vs official 4.1163 % for 2023 — consistent (within normal rounding).",
    "citation": {
      "source": "World Bank",
      "series_id": "FP.CPI.TOTL.ZG",
      "source_url": "https://data.worldbank.org/indicator/FP.CPI.TOTL.ZG?locations=USA",
      "license": "CC BY 4.0",
      "citation_text": "World Bank, World Development Indicators, series FP.CPI.TOTL.ZG (Inflation, consumer prices, annual %) …"
    }
  }
}

Batch mode

Pass items (array of parameter objects, each may set its own operation) to run many queries in one call — one pay-per-event charge per successful item; failures are free and reported per item.

{
  "operation": "indicator",
  "items": [
    { "operation": "verify", "indicator": "gdp_growth", "country": "Guyana", "period": "2022", "claimed_value": 62.3 },
    { "operation": "indicator", "indicator": "govt_debt_gdp", "country": "Japan", "latest_only": true },
    { "operation": "fx", "amount": 1000, "from": "USD", "to": "JMD", "date": "2023" }
  ]
}

Batch runs and upstream fetches

Each batch item runs with a fresh internal context, so upstream fetches are not memoized across items. This is deliberate: the memo caches failures for its lifetime (by design, to protect the per-request budget), and a shared context would let one transient IMF failure serve degraded fallback data — while still charging lookups — for the remainder of a run. The cost is some repeated upstream fetches on large batches; correctness of paid results wins.

Pricing

Pay per successful query event — failures are never charged. Two event tiers:

Event	Operations	Why
statcite-lookup	indicator, series, snapshot, inflation, fx, search, registry	A cited value lookup — cheap, high-volume
statcite-verify	verify	Checks a claimed figure against the official series with diagnostics — the differentiated capability, priced higher

A generous free web tier of the same engine exists at statcite.com (MCP + REST); this actor is for pipelines that want Apify-managed runs, storage, scheduling, and invoicing.

Use from AI agents (MCP)

Via the Apify MCP server, agents can discover and call this actor directly. There is also a free hosted MCP endpoint at https://statcite.com/mcp.

Support

hello@statcite.com · statcite.com/docs.html