How to Scrape CPSC Product Recalls

Consumer product recalls arrive as detailed notices, but commerce teams usually need structured fields that can be matched against catalogs and suppliers. The CPSC Product Recalls Scraper converts the official public feed into rows containing products, hazards, remedies, companies, dates, images, and source links.

Search the recall feed

Use terms that identify a brand, product family, or hazard:

{"queries":["lithium-ion battery","stroller"],"startDate":"2024-01-01","maxResultsPerQuery":100}

Each row preserves the recall ID and title, recall date, description, consumer contact, countries, product details, hazard and remedy text, incidents and injuries, manufacturer or retailer records, images, and the official notice URL when available. Nested source fields are normalized into arrays so downstream systems do not need to parse the original response shape.

Deduplicate on recall ID. The same notice may match a brand query and a hazard query, while one recall can cover several product variants or companies. Keep the matched query as provenance, but maintain one canonical recall record and separate match relationships downstream.

Date filters help constrain monitoring windows, yet older recalls still matter for long-lived inventory and resale listings. Run a recent window frequently and a broader historical sweep less often. Normalize brand and company names for matching without discarding the source wording. Product descriptions can be broad, so require a human review or strong identifier match before suppressing a listing.

The feed is an intake source, not a substitute for the current recall notice. Product scope, model numbers, remedy instructions, and consumer contacts must be confirmed from the official page before customer communication or enforcement. Record the collection time, retain the CPSC link, and make review outcomes auditable. That combination turns public recall notices into a practical safety workflow without overstating what keyword matching can prove.