Economic Data Tools
Every other tool family here prices an instrument. These two read the economy the instrument trades in — inflation, unemployment, GDP, policy rates, the Treasury curve, money supply, sentiment, housing.
| Tool | Purpose |
|---|---|
economicSeries | History for one indicator, with optional year-over-year or period-over-period transform |
economicRelease | The latest print for up to eight indicators at once, with change, year-over-year move and staleness |
No API key
The upstream is DBnomics, a free aggregator run by CEPREMAP that re-serves FRED, BLS, Eurostat, ECB, IMF, OECD and a long tail of national statistics offices behind one JSON API. It needs no key and no registration, which is the reason it was chosen: FRED's catalogue is reachable without a FRED key, so every upstream in this server stays free.
There is no configuration. There is also no published rate limit, which is not the same as there being none — responses are cached for an hour, because the data underneath moves monthly at best.
Two ways to name a series
A catalogue id. indicator: "us-core-pce" — a curated shortlist of the series people actually ask for, because nothing knows offhand that core PCE is PCEPILFE. Each entry carries its unit, its frequency, and an opinion about how the number should be read.
A raw path. seriesId: "Eurostat/prc_hicp_midx/M.I15.CP00.EA" — any series DBnomics holds, addressed as provider/dataset/code. The catalogue is a convenience, never a limit; browse identifiers at db.nomics.world.
The catalogue is US-first because that is what "key economic data" usually means in practice. Everything else goes through seriesId.
Index series are returned as a change, not a level
"CPI is 320.4" answers no question anyone asked. Series marked as an index therefore default to a year-over-year transform, while rates and levels are reported exactly as published. Override with transform, which takes none, pop_pct (period-over-period) or yoy_pct.
Changes are matched by date, never by counting back N rows. A daily yield series skips weekends and holidays, a monthly series can carry a suspended period, and a survey can be discontinued and resume — indexing back twelve rows to find "a year ago" is right only for a series with no gaps.
Read staleDays before calling a figure current
Economic data lags the period it measures: a CPI print lands weeks after the month it describes, and a quarterly GDP revision months after the quarter. Every economicRelease row reports how many days old the latest reading is, so a three-month-old print is never mistaken for today's inflation.
economicRelease also survives partial failure: one indicator failing leaves the rest of the snapshot intact, with the error reported on that row.