A search engine for the world’s statistics.

25,798 datasets from 394 official sources. Pick a bundle or clone the entire catalog locally. It stays synced as sources publish updates.

$pip install subsetsio

Free forever · no account · works offline.

Prefer not to run anything? Get a hosted key: 500 free queries, no card →

World BankIMFECBWHOUNHCRCDCBundesbankAus. Bureau of Statistics
+ 386 more official sources →

Don’t quote the number. Query it.

Three tools: search, info, and query. Every result includes its source, license, and last-updated date.

$ subsets sync --bundle small   # the default: ~3,500 tables, ~200 MB
✓ synced — your copy, and it works offline.

$ subsets search "malaysia inflation"
1. dosm-cpi-headline-inflation   Monthly Headline CPI Inflation · DOSM (Malaysia)
2. imf-cpi   Consumer Price Index (CPI) · IMF

$ subsets query 'SELECT date, inflation_yoy FROM "dosm-cpi-headline-inflation" LIMIT 3'
┌────────────┬───────────────┐
│ date       │ inflation_yoy │
├────────────┼───────────────┤
│ 2026-04    │ 1.4           │
│ 2026-05    │ 1.2           │
│ 2026-06    │ 1.1           │
└────────────┴───────────────┘
source: DOSM (Malaysia) · CC BY 4.0 · updated 2026-07-18

Bundles control how much of the catalog you clone: small by default, medium adds the larger sources, and large takes everything we bundle today. Re-syncing downloads only what changed. How bundles work →

Run it locally, or let us run it

Local free forever

Full catalog available, full search, no account. Each source keeps its own license; the pipelines are MIT, so you can run the whole system yourself.

$pip install subsetsio

Hosted nothing to install

The same catalog and search engine, as an HTTP API. No downloads, no syncing, no local mirror to manage. Built for apps and agents in production.

Get a key500 free queries, then from $19/mo