Canada Product Recalls & Safety Alerts

Scrape Canada's official product recall and safety alert database — Health Canada's recalls-rappels.canada.ca (open.canada.ca) — and get structured JSON recall records: consumer products, vehicles, food, drugs and health products. Filter by keyword, category and last-updated date.

Built for compliance teams, importers, retailers and safety monitors who need a reliable, always-current feed of Canadian recall activity — no API keys, no login, no proxy budget.

Why this Actor

• One source, one job — reads the Government of Canada recall database directly; no third-party aggregators, no licensing ambiguity.
• No API key, no login, no proxy — the site is fully public; the Actor just needs a browser User-Agent (the site 418s plain script clients).
• Structured, reference-shaped output — every recall is one JSON row with 17 fields matching the community dromb/canada-recalls-safety-alerts schema, so existing consumers can switch with zero changes.
• Filters built in — keyword full-text search plus post-filters on category and last-updated date.
• Cheap and fast — pay only per recall extracted, plus Apify's standard platform events.

How it works

The Actor runs a full-text search against the site's Drupal search index (/en/search?search_api_fulltext=<keyword>), pages through the results, visits each recall page (/en/alert-recall/<slug>), and parses the structured Drupal fields into clean records. Archived recalls are excluded (matching the reference actor's behaviour), so every row is an active recall.

Input

{
  "keyword": "battery",
  "category": "Electronics",
  "dateFrom": "2026-06-01",
  "maxItems": 50
}

Field	Type	Description
keyword	string	Full-text search term (e.g. battery, croissant, vehicle). Empty = browse the latest recalls.
category	string	Category filter — case-insensitive partial match against the recall category (e.g. Electronics, Household items, Vehicles).
dateFrom	string	Only include recalls last updated on or after this date (YYYY-MM-DD).
maxItems	integer	Maximum recall records to extract (default 50, max 500).

All fields are optional — {} returns the most recent active recalls.

Output (one item per recall)

{
  "recallId": 82426,
  "title": "Portable Clip-On Rechargeable Mini Fan recalled due to fire hazard",
  "dateUpdated": "2026-08-12",
  "recallClass": null,
  "category": "Household items",
  "summary": "Fire hazard",
  "affectedProducts": "Portable Clip-On Rechargeable Mini Fan",
  "companies": "Consumer product safety",
  "hazard": null,
  "issue": "Fire hazard",
  "whatToDo": "Consumers should immediately stop using the recalled product and return it to CTG Brands Inc. for a refund.For more information, consumers can contact CTG Brands Inc. by telephone at 1-888-781-8882, 8:30 a.m. to 5:00 p.m. EST, Monday to Friday, or visit the company website...",
  "archived": false,
  "status": "active",
  "language": "en",
  "sourceUrl": "https://recalls-rappels.canada.ca/en/alert-recall/portable-clip-rechargeable-mini-fan-recalled-due-fire-hazard",
  "dataSource": "open.canada.ca",
  "rawSourceType": "recall_record"
}

Field notes (matching the reference actor): category strips the top-level taxonomy prefix (Consumer products - Electronics → Electronics); summary/issue mirror the short issue label (last - segment); whatToDo is the long-form consumer guidance (null when the page has none); recallClass is Type I/II/III for health-product recalls, Class 1/2/3 for food recalls, else null; companies is the issuing organization from publisher metadata (Consumer product safety, CFIA, TC); archived is always false / status always active because the search surface excludes archived recalls.

Use cases

• 🛡️ Compliance monitoring — weekly cron on a product category (electronics, toys, food) and alert on new recalls via Telegram or webhook.
• 🏭 Importer / retailer screening — check a product line against active recalls before ordering or listing.
• 📊 Safety trend analysis — pull all recalls in a category over a window and analyse hazard mix, issuing agency and product types.
• 🗄️ Data enrichment — join recall IDs against internal product SKUs for a live exposure report.

Run it from the API

Trigger the Actor and get results in one call:

curl -X POST "https://api.apify.com/v2/acts/darknezz~canada-recalls-safety-alerts/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keyword":"battery","category":"Electronics","dateFrom":"2026-06-01"}'

Python (Apify SDK):

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("darknezz~canada-recalls-safety-alerts").call(
    run_input={"keyword": "battery", "category": "Electronics", "dateFrom": "2026-06-01"},
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["recallId"], item["title"])

Scheduling: attach an Apify Schedule (daily or weekly) to keep a category watchlist topped up — the dateFrom filter plus a stored last-run date gives you an incremental feed of new recalls only.

Limitations

• HTML scraping — the site has no public JSON API, so parsing is tied to the Drupal markup. The parser targets stable field machine-names; a site redesign would need a rebuild.
• English only — records are extracted from the /en/ surface; French (/fr/) records are not covered.
• Archived recalls excluded — matching the reference actor; historical/archived records are not returned.
• Politeness — detail fetches are paced; very large result sets (500+ items) take proportionally longer.

Pricing

Pay per event: a small fee for each recall extracted, plus Apify's standard platform events (actor start + dataset item). No subscription — you pay only for what you run.

FAQ

Do I need an API key or government login? No — the recall database is fully public.

Why do I get nothing for some keyword/category combos? The category filter is a post-filter on the parsed category. If the search finds no recalls whose category matches, the dataset is empty — broaden the keyword or drop the category.

How fresh is the data? The site is updated continuously; dateUpdated is the page's last-modified date, so dateFrom gives you an incremental window.

Can I get archived recalls? No — the search surface excludes them, and the Actor follows that (as does the reference actor).

Is scraping this legal? The data is Canadian government open information published for public consumption; you are responsible for how you use it.