FDA Data Scraper (openFDA)

Pull FDA data from the official openFDA API — drug, device, and food recalls, adverse events, and drug labels — into one unified, clean, AI-ready JSON dataset. No API key required. Most FDA scrapers cover a single dataset; this one covers six, with consistent fields and a simple keyword search across all of them.

What does this Actor do?

Select one or more FDA datasets and get normalized records back:

• 💊 Drug recalls (enforcement reports) — firm, product, reason, class, date
• 🍎 Food recalls (enforcement reports)
• 🩺 Device recalls
• ⚠️ Drug adverse events (FAERS) — product, reaction, seriousness, death flag
• 🔬 Device adverse events (MAUDE)
• 🏷️ Drug labels (SPL) — brand, generic, manufacturer, purpose
• 🔎 Unified search — one keyword or a raw openFDA query applied across every dataset, plus date filtering
• 🤖 AI-ready JSON — normalized top-level fields plus the full raw record, built for dashboards, monitoring, and AI agents (via the Apify MCP server)

Why scrape openFDA?

FDA data powers real decisions across industries, but the raw API is spread across many endpoints with different schemas. Pharma companies, health-tech, medical-device makers, product-liability lawyers, insurers, supply-chain teams, and journalists use this Actor to:

• Monitor recalls for their products or competitors in real time
• Track drug/device safety signals (adverse events) for pharmacovigilance
• Build compliance and risk dashboards
• Feed regulatory data into models and AI agents

Input example

{
  "datasets": ["drugRecalls", "foodRecalls", "deviceRecalls"],
  "searchTerm": "insulin",
  "dateFrom": "2025-01-01",
  "maxRecordsPerDataset": 500
}

• searchTerm — simple keyword matched across firm/product/reason (easiest).
• searchQuery — raw openFDA expression for power users, e.g. classification:"Class I" or recalling_firm:Pfizer.
• datasets — mix and match any of the six.

Output example

{
  "dataset": "drugRecalls",
  "id": "D-0626-2026",
  "date": "2026-07-01",
  "firm": "Annora Pharma Private Limited",
  "product": "Metformin HCl Tablets, 500 mg",
  "reason": "Presence of foreign tablets: possible product mix-up",
  "classification": "Class II",
  "status": "Ongoing",
  "raw": { "…": "full openFDA record" }
}

Every record keeps the complete original openFDA payload under raw, so no data is lost.

How much does it cost?

Pay per record returned — a small fee each. Pulling 1,000 records costs a fraction of a dollar. No subscription.

FAQ

Do I need an API key? No. It works key-free (1,000 requests/day). For large jobs, add a free openFDA key (apiKey input) to raise the limit to 120,000/day — get one at open.fda.gov/apis/authentication.

Is this affiliated with the FDA? No. It uses the FDA's official public openFDA API. Not endorsed by the FDA. openFDA data has usage caveats (do not use to make clinical decisions) — see the FDA's terms.

Can AI agents use it? Yes — works with the Apify MCP server, so Claude, ChatGPT, and other agents can call it and pay per use.

Support

Need another FDA dataset (NDC directory, drug shortages, 510k clearances) or a feature? Open an issue in the Issues tab — actively maintained, answered within one business day.