openFDA Scraper

Pull structured public‑health data straight from the U.S. FDA's official openFDA API — no scraping, no anti‑bot fights, no auth required. One Actor covers 18 official FDA datasets with the full openFDA query grammar (search, count, sort, pagination), and returns clean, table‑ready records with the complete original object preserved.

Great for pharmacovigilance, medical‑device safety monitoring, recall tracking, regulatory & competitive intelligence, research, and data journalism.

Datasets covered

Group	Endpoint	What it is
Drug	drug/event	Adverse‑event reports (FAERS)
Drug	drug/label	Structured product labeling (SPL)
Drug	drug/enforcement	Drug recalls / enforcement reports
Drug	drug/ndc	National Drug Code directory
Drug	drug/drugsfda	Drugs@FDA approvals
Device	device/event	Medical‑device adverse events (MAUDE)
Device	device/510k	510(k) clearances
Device	device/classification	Device classification
Device	device/enforcement	Device recalls / enforcement
Device	device/recall	Device recalls
Device	device/pma	Pre‑market approvals
Device	device/udi	Unique device identifiers (GUDID)
Device	device/registrationlisting	Registration & listing
Device	device/covid19serology	COVID‑19 serology test performance
Food	food/enforcement	Food recalls / enforcement
Food	food/event	Food & cosmetic adverse events (CAERS)
Animal	animalandveterinary/event	Animal & veterinary adverse events
Tobacco	tobacco/problem	Tobacco problem reports

Input

Field	Type	Description
endpoint	select	Which openFDA dataset to query (default drug/event).
search	string	openFDA search query, e.g. patient.drug.medicinalproduct:aspirin or classification:"Class I" AND report_date:[20240101 TO 20241231]. Empty = most recent records.
count	string	Aggregation field → returns {term, count} buckets instead of records, e.g. patient.reaction.reactionmeddrapt.exact. Append .exact to count whole phrases.
sort	string	e.g. receivedate:desc, decision_date:desc (record mode only).
maxResults	integer	Max records/buckets to return (auto‑paginated). Default 100.
skip	integer	Starting offset (record mode; openFDA caps skip at 25000).
apiKey	string	Optional free openFDA API key — raises the rate limit. Not required.

See each endpoint's searchable fields in the openFDA API reference.

Output

Because the 18 datasets have very different native shapes, every record carries a small set of normalized convenience columns — endpoint, recordId, primaryDate, name, company, classification, reason, status, country, state — plus the full, lossless original record under raw, so nothing from the source is dropped.

In count mode each record is a {term, count} bucket.

Example (drug recall):

{
  "endpoint": "drug/enforcement",
  "mode": "search",
  "recordId": "D-0595-2026",
  "primaryDate": "2026-06-04",
  "name": "Gas-X Extra Strength, SIMETHICONE 125 mg ...",
  "company": "Haleon US Holdings LLC",
  "classification": "Class I",
  "reason": "Chemical Contamination ...",
  "status": "Ongoing",
  "country": "United States",
  "state": "NJ",
  "raw": { "...": "the complete original openFDA object" }
}

Pricing

Pay‑per‑result: you are charged per record (or count bucket) returned. Failed or empty queries cost nothing.

Notes

• Data comes directly from api.fda.gov — the official FDA endpoint. This Actor does not modify or validate the underlying data; consult FDA's disclaimer. Do not use openFDA data to make decisions about medical care.
• openFDA is public and free; an optional API key only raises rate limits.
• Deep paging is capped at skip=25000 by openFDA — for very large pulls, window through the data with a date‑range search + sort.