EU Safety Gate Product Recall Matcher

Screen your product catalog against recent EU Safety Gate alerts and receive evidence-backed candidates for human review. The Actor compares GTINs, brands, models, and product names, labels every match with a confidence level, and links directly to the official alert.

Built for marketplaces, retailers, importers, distributors, product-safety teams, insurers, and compliance operations that need repeatable recall monitoring without manually searching weekly publications.

What it does

• Reads non-food product alerts from the official EU Safety Gate API and its weekly reports (full history back to 2005).
• Accepts pasted catalog products or a public Google Merchant product feed.
• Matches exact GTINs, brand plus model, or broader brand plus product-name evidence.
• Keeps match confidence and matching evidence visible in every row.
• Links every candidate to the official Safety Gate alert for verification.
• Produces a run summary with products checked, alerts read, matches by confidence, and products lacking usable identifiers.
• Charges per product screened, never per match found.

Quick start

Provide catalog records with as many identifiers as possible.

{
  "products": [
    {
      "id": "SKU-1",
      "brand": "Scheppach",
      "model": "HM90SL",
      "name": "Sliding compound mitre saw"
    }
  ],
  "weeksBack": 4,
  "minConfidence": "probable",
  "maxProducts": 5000
}

Add a GTIN whenever available. An exact GTIN creates the strongest candidate and avoids the ambiguity of free-text model and product names.

Match confidence

Confidence	Rule	Recommended use
certain	Exact normalized GTIN match	Highest-priority review
probable	Brand and model both match	Review against the official alert
possible	Brand matches and product names strongly overlap	Broad monitoring and manual triage

A brand alone is never considered a match. The Actor does not choose a business action, remove a listing, or claim legal certainty. Even exact identifier matches should be verified against the official alert, the physical product, supplier records, and your internal procedures.

Input

Field	Description
products	Catalog objects containing id, brand, name, model, and GTIN where available.
feedUrl	Public Google Merchant feed URL used when products is empty.
weeksBack	Number of weeks of Safety Gate alerts to read. Default: 4.
minConfidence	Lowest confidence included: certain, probable, or possible.
maxProducts	Maximum catalog products screened in the run.

A product with neither a GTIN nor a brand cannot be matched. The Actor counts such products in productsWithoutIdentifiers rather than presenting them as cleared.

Output

The dataset contains two row types:

• match — a free candidate containing your product identifiers, confidence, matching evidence, official alert number, date, risk information, notifying country, and source URL.
• summary — a free run-level record with products checked, alerts read, the date window, match counts, and unmatchable-product count.

Matches are intentionally free. Billing per match would create the wrong incentive for a safety-monitoring product; the billable unit is the catalog product that was screened.

Use the dataset view selector to switch between Matches and Run summary, then export JSON, CSV, Excel, XML, or HTML.

Pricing

This Actor uses pay-per-event pricing with platform usage included:

• $0.01 per run start at the default 512 MB memory.
• $0.0005 per product checked — $0.50 per 1,000 catalog products.
• Matches, summaries, unmatchable products, and filtered alerts are free.

Example totals at default memory: 100 products cost about $0.06, 1,000 cost $0.51, and 10,000 cost $5.01. Users can set a maximum charge per run; the Actor honors that limit and stops early.

Data source and reliability

Safety Gate is the European Union rapid alert system for dangerous non-food products. The Actor reads a published data feed and attaches the official alert page to each candidate. If the feed stops answering partway through, the run log states how many alerts were actually read instead of silently presenting a partial window as complete.

Alerts come only from official Safety Gate sources: the JSON API the official website uses (roughly the most recent month, alert by alert) plus the official weekly reports, which cover every week back to 2005. Alerts found in both are counted once. If part of your window is covered by neither source, every row carries a sourceWarning (and the log states the gap), and sourceLastUpdated shows the newest alert date the data holds. A missing match is then not proof of safety. When no alerts can be read at all, the run fails without charging per-product fees; when zero alerts fall in the window, no per-product fee is charged either.

The alert window is capped at 6,000 records per run. Use a narrower weeksBack value for routine schedules and a wider one for occasional historical checks.

Important limitations

This Actor is an automated screening and decision-support tool. False positives and false negatives are possible. It is not legal, regulatory, product-safety, medical, or compliance advice, and it does not certify that a catalog is recall-free. Safety Gate does not cover food, medicines, or medical devices. Public records may be incomplete, revised, or delayed.

Always verify candidates on the official alert page and follow qualified legal, regulatory, and product-safety guidance before taking action. Use the data responsibly and comply with applicable laws, contracts, and source terms.

Automation

Schedule weekly or daily runs in Apify Console, trigger the Actor by API, or connect results to Slack, email, Google Sheets, Make, Zapier, webhooks, or an internal case-management system. Store the alert number and product ID to deduplicate repeated candidates across runs.

FAQ

Why did a product produce no match?

No relevant alert may exist in the selected window, the product may lack identifiers, or the source text may not align strongly enough with your selected confidence level. A missing match is not proof that the product is safe or recall-free.

Why are match rows free?

You pay for each product screened. Matches and summaries are free so pricing cannot reward or discourage a positive result.

Where can I get help?

Open the Actor's Issues tab with the run ID, alert number, confidence level, and a redacted example product. Do not include private feeds, credentials, or personal data.

Related Actors

• Google Shopping Feed Checker
• E-commerce Product Scraper
• Dataset Cleaner & Deduplicator