DEV Community

Noble Ronin
Noble Ronin

Posted on

FDA Recalls Have a Free JSON API — Every Drug, Device and Food Recall Without Scraping

#api #healthcare #webscraping #data

FDA Recalls Have a Free JSON API

If you need FDA recall data — which drugs, devices and foods got pulled, why, by whom, and how serious it was — you do not need to scrape fda.gov, and you do not need a paid API. The FDA runs openFDA, a free, keyless JSON API over its own enforcement, adverse-event and labeling databases.

No sign-up. No API key. No token. There's a higher rate limit if you register a free key, but every endpoint below works with nothing at all.

The base 

https://api.fda.gov
Enter fullscreen mode Exit fullscreen mode

Recalls live under the enforcement endpoints, one per product area:

curl "https://api.fda.gov/drug/enforcement.json?limit=1"
curl "https://api.fda.gov/device/enforcement.json?limit=1"
curl "https://api.fda.gov/food/enforcement.json?limit=1"
Enter fullscreen mode Exit fullscreen mode

Each response has a meta block (with results.total) and a results array. A single drug-recall record carries recalling_firm, product_description, reason_for_recall, classification (Class I / II / III — I is the most serious), status (Ongoing / Completed / Terminated), recall_initiation_date, distribution_pattern, code_info, plus an openfda block that links the product back to NDC codes and brand/generic names.

Search with the search parameter

openFDA uses a simple field:value query syntax. Quote multi-word values and URL-encode:

# All Class I (most serious) drug recalls
curl 'https://api.fda.gov/drug/enforcement.json?search=classification:"Class+I"&limit=5'

# Recalls that are still Ongoing
curl 'https://api.fda.gov/drug/enforcement.json?search=status:"Ongoing"'

# Ranges and booleans work too — food recalls initiated in 2026
curl 'https://api.fda.gov/food/enforcement.json?search=recall_initiation_date:[20260101+TO+20261231]'
Enter fullscreen mode Exit fullscreen mode

At the time of writing there are 1,733 Class I drug recalls on record and 2,580 with an Ongoing status — all as structured JSON, no scraping.

Every drug, device and food recall as JSON

Count instead of list

Add count=<field>.exact and openFDA returns an aggregation instead of raw rows — great for dashboards:

curl "https://api.fda.gov/drug/enforcement.json?count=classification.exact"
Enter fullscreen mode Exit fullscreen mode

Right now that returns Class II = 14,381, Class I = 1,733, Class III = 1,701. One call, the whole distribution.

Page through everything 

curl "https://api.fda.gov/device/enforcement.json?limit=100&skip=200"
Enter fullscreen mode Exit fullscreen mode

limit (max 1000) and skip page the results. The device enforcement set alone is 39,539 records; food is 29,238.

Beyond recalls

Same base, same keyless deal, for two more datasets people usually ask about in the same breath:

# Adverse event reports (FAERS)
curl "https://api.fda.gov/drug/event.json?limit=1"

# Structured drug product labeling
curl "https://api.fda.gov/drug/label.json?limit=1"
Enter fullscreen mode Exit fullscreen mode

The only two rules

  1. Be reasonable with volume — anonymous calls are rate-limited (a registered free key just raises the ceiling).
  2. Page with limit + skip, and remember openFDA data is a snapshot the FDA refreshes periodically, not a live wire.

That's the whole thing.

If you'd rather skip the paging, the search encoding and the field-mapping, the FDA Recalls Scraper on Apify wraps exactly these endpoints — drug/device/food, class and status filters in, structured rows out.

📌 Full endpoint cheatsheet (copy-paste reference): github.com/noble-ronin/fda-recalls-api

Top comments (0)