| Dataset | FRED series with ALFRED vintages |
|---|---|
| Dataset id | fred.alfred |
| Series | 0 |
| Observations | 0 |
| Countries | 0 |
| Period range | none |
| Vintages | 0 |
| Last fetched | never |
| Poll interval | synced, not polled |
| Archive | fred/alfred/*.json.gz |
| Licence | FRED Terms of Use |
| Redistributable | restricted |
| Terms read by a person | 2026-09-11 |
| Attribution | Source: FRED, Federal Reserve Bank of St. Louis |
The connector's own docstring, verbatim. It lives beside the code so it cannot drift from it.
FRED / ALFRED connector (St. Louis Fed).
FRED is unusual and that is exactly why it is here. Most sources only tell you what the data says today, so revisions are learned by polling and waiting. ALFRED publishes the whole revision history, so every past vintage can simply be asked for -- which makes this the one source that can validate our point-in-time queries against published ground truth rather than a test we wrote ourselves.
Its observations endpoint, given a wide realtime range, returns one row per (period, vintage) with the realtime window in which that value was current. That is already a delta stream: a row's realtime_start IS the vintage where the value first appeared. Grouping by realtime_start reconstructs the real vintages, which are then replayed in order by fred_sync.py.
Requires a free API key from https://fredaccount.stlouisfed.org/apikeys, set as FRED_API_KEY in .env.