Wikimedia Wikidata Query Service

Use in your agent

Query Wikidata SPARQL into pandas and Parquet. Checked against the provider's documentation on Nov 2, 2026.

Access

• No account or key. Send a descriptive User-Agent and use the public SPARQL endpoint.
• JSON is selected with format=json or the application/sparql-results+json Accept header.

Request

curl --get 'https://query.wikidata.org/sparql' -H 'User-Agent: SignalLabCheck/1.0 (https://signal-lab-jet.vercel.app)' --data-urlencode 'query=SELECT ?item ?itemLabel WHERE { VALUES ?item { wd:Q16 wd:Q30 wd:Q145 } SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } }' --data 'format=json'

Format

SPARQL Results JSON with head.vars and results.bindings. Each binding describes a value's type, value and optional language or datatype.

Load into pandas and Parquet

Needs Python 3 with requests, pandas and pyarrow.

"""Three Wikidata entities from WDQS -> pandas -> Parquet."""
import requests
import pandas as pd

URL = "https://query.wikidata.org/sparql"
QUERY = """
SELECT ?item ?itemLabel WHERE {
  VALUES ?item { wd:Q16 wd:Q30 wd:Q145 }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}
"""
headers = {"User-Agent": "SignalLabCheck/1.0 (https://signal-lab-jet.vercel.app)"}
r = requests.get(URL, params={"query": QUERY, "format": "json"}, headers=headers, timeout=60)
r.raise_for_status()
bindings = r.json()["results"]["bindings"]
df = pd.DataFrame([{k: v.get("value") for k, v in row.items()} for row in bindings])
df.to_parquet("wikidata_query_sample.parquet", index=False)
print(len(df), "entities; labels:", ", ".join(sorted(df["itemLabel"])))

Run by us on 2026-11-02 (python 3.13.5 | pandas 3.1.0 | pyarrow 26.0.0 | requests 2.34.2). Result: 3 entities; labels: Canada, United Kingdom, United States

Limits

• Query timeout: 60 seconds.
• Processing allowance: 60 seconds per 60 seconds for one user-agent-and-IP client.
• Error allowance: 30 erroneous queries per minute; clients exceeding limits are throttled.

Quirks

• Flatten each SPARQL binding's nested value object before constructing a table.
• The query-service dataset does not exactly match the RDF dump, and asynchronous updates introduce lag.
• The main and scholarly graphs are split across separate public endpoints; choose the graph that matches the topic.

Terms and attribution

• Wikidata structured data in the main data namespaces is dedicated under CC0 1.0.
• Text outside the structured-data namespaces is available under CC BY-SA and may have additional terms.
• Follow Wikimedia's User-Agent, robot and API etiquette policies when automating requests.

Sources

• https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service: “Run, test, download the data as XML, JSON, CSV, or TSV.”
• https://www.mediawiki.org/wiki/Wikidata_Query_Service/Implementation: “Query timeout of 60 seconds”
• https://www.mediawiki.org/wiki/Wikidata_Query_Service/Implementation: “One client (user agent + IP) is allowed 60 seconds of processing time each 60 seconds”
• https://www.wikidata.org/wiki/Wikidata:Copyright: “Creative Commons CC0 License”