DBnomics

What it is

DBnomics puts one API in front of 93 economic data providers — INSEE, the ECB, the Federal Reserve Board, the Bank of England, the IMF, the World Trade Organization, Statistics Japan — and keeps each provider's own codes rather than inventing a house schema. A series is addressed as provider, dataset and series code, so the daily dollar–euro reference rate is ECB/EXR/D.USD.EUR.SP00.A in the URL, in the Python client and in the CSV alike. A lookup you worked out on the ECB's own site transfers unchanged.

It is operated by CEPREMAP, a French public economic research centre, and co-funded by partner institutions including the Banque de France, the Agence française de développement, France Stratégie and the Direction générale du Trésor. That is the useful thing to know: this is a publicly funded research utility that has been running since 2015, not a company with a runway.

Pricing

Free, and the API takes no key and no account: an unauthenticated request for a series returns data, and the OpenAPI specification defines no authentication scheme at all. The website's own download button is the exception — it requires a free account. The free tier is capped by a fair-use limit on API calls that the sponsorship page describes but does not quantify.

The two paid tiers are sponsorships rather than product tiers. Supporter is 1,500 euros a year and Sustainer 15,000 euros a year; both remove the request cap, and Sustainer adds priority support and the right to request up to two new data sources annually. Neither unlocks data that the free tier cannot reach.

Data & coverage

93 providers and 47,062 datasets, both counted from the API on 14.09.2026; the same response puts the series total at 1.73 billion. Frequencies run from annual to daily: the ECB daily reference rate carries 7,148 observations back to 04.01.1999, the Federal Reserve H.15 dataset holds 124 interest-rate series, and the IMF primary commodity price dataset holds 1,236.

Freshness depends on the fetcher behind each provider. The documentation says each fetcher runs once a day by default and that new data should appear within 24 hours under normal circumstances. Checked on 14.09.2026, the ECB daily rate ran through 04.09.2026.

Integrations

The REST API answers JSON by default and CSV with a format parameter, so a series lands in a spreadsheet without a client at all. The Python package dbnomics is the first-party client — published from a nomics.world address, AGPL-3.0-or-later, version 1.2.7 released 18.06.2025 — and its fetch_series returns a pandas DataFrame.

Everything else is community work, and the documentation labels it as such: rdbnomics for R, DBnomics.jl for Julia, plus Stata, EViews, Gretl and Matlab clients. A local instance can be run from the project's dbnomics-docker images and Compose files.

Limitations

It is not a market data API. No equities, no options, no order books, no company fundamentals. If you arrived here from a stock screener, this is the wrong shelf.

The licence on the code is not the licence on the data. Each source repository carries its own licence and most of the fetchers are AGPLv3+, while the aggregated datasets are published under the Open Database Licence — but the legal terms also state that data remains subject to the terms of its original source provider. Those two lines sit next to each other in the same four-point terms and they do not resolve each other. Before republishing anything, check the provider. DBnomics does not carry paid or non-redistributable data at all, which is why the catalogue is official statistics rather than vendor feeds.

Fetchers break, and the project says so. Its own FAQ notes that many fetchers sit in error status on the dashboard, because a fetcher is only as stable as the provider website it scrapes. A quiet provider is indistinguishable from a broken fetcher until you check the last observation.

The API is built for focused queries, not bulk, and the documentation says so. The website download button stops at 1,000 series, and the Python client returns 50 unless you pass max_nb_series explicitly.

Only the Python client is maintained by the project. The R client is the one most economists will reach for, and its CRAN release is 0.6.4 from 25.10.2020. Source for everything lives on the project's own GitLab rather than GitHub, which makes casual contribution and mirroring harder than the licence implies.

Alternatives

Nasdaq Data Link is where the commercial descendants of the old Quandl economic catalogue ended up, sold rather than published. OpenBB is the other shape of the same idea — one Python interface over many providers, including several of the same public sources, rather than one converted copy of them.