Banque de France Webstat API

Use in your agent

Load a Webstat series into pandas and Parquet. Checked against the provider's documentation on Nov 1, 2026.

Access

• Create a free Webstat account and API key.
• Set the key in BANQUE_DE_FRANCE_API_KEY; send it as Authorization: Apikey <key>.

Request

curl -H "Authorization: Apikey $BANQUE_DE_FRANCE_API_KEY" 'https://webstat.banque-france.fr/api/explore/v2.1/catalog/datasets/observations/exports/json?select=time_period_end,obs_value&where=series_key%3D%27MIR1.M.FR.B.L23FRLA.D.R.A.2230U6.EUR.O%27'

Format

JSON array of selected observation fields; the full observations table includes dataset and series IDs, French and English titles, period start and end, value, status, update timestamp and observation attributes.

Load into pandas and Parquet

Needs Python 3 with requests, pandas and pyarrow.

"""Banque de France Webstat series -> pandas -> Parquet. Needs a free API key."""
import os
import requests
import pandas as pd

URL = "https://webstat.banque-france.fr/api/explore/v2.1/catalog/datasets/observations/exports/json"
key = os.environ["BANQUE_DE_FRANCE_API_KEY"]
params = {
    "select": "time_period_end,obs_value",
    "where": "series_key='MIR1.M.FR.B.L23FRLA.D.R.A.2230U6.EUR.O'",
}
headers = {"Authorization": f"Apikey {key}", "Accept": "application/json"}
r = requests.get(URL, params=params, headers=headers, timeout=60)
r.raise_for_status()
df = pd.DataFrame(r.json())
df["time_period_end"] = pd.to_datetime(df["time_period_end"])
df.to_parquet("banque_de_france_webstat.parquet", index=False)
print(len(df), "observations", df["time_period_end"].min().date(), "to", df["time_period_end"].max().date())

Not run: it needs a free API key.

Limits

• Daily quota: 10,000 requests per user account.
• The API returns HTTP 429 after the daily quota is reached.
• Updates are published three times daily at 09:00, 13:00 and 21:00 Paris time.

Quirks

• Use a where or refine clause on observations; the guide warns that omitting it can return all rows.
• The API does not join its four main catalog tables, so request series and codelist metadata separately.
• The provider asks clients to store downloaded data locally instead of polling repeatedly.

Terms and attribution

• Cite the source of any reused dataset.
• Check the open or particular licence attached to each dataset before download or extraction.
• Creating multiple accounts to bypass the daily quota is prohibited.

Sources

• https://webstat.banque-france.fr/pages/guide-migration-api/?flg=en-us: “Using the API requires a key that can be passed as an HTTP parameter in the URL, or, and this is the recommended way, in an HTTP header.”
• https://webstat.banque-france.fr/pages/guide-migration-api/?flg=en-us: “Daily Quota: 10,000 Requests”
• https://webstat.banque-france.fr/terms/terms-and-conditions/?flg=fr-fr: “En toute hypothèse, mentionner la source des JEUX DE DONNEES en cas de réutilisation de tout ou partie de ceux-ci.”