One API for every US product recall.
Normalized data across CPSC, FDA, NHTSA, and USDA. Real-time webhooks, historical depth, and AI-enriched fields like severity and affected groups — so your platform can act on a recall before customers do.
What's in the API
The same recall data the four federal agencies publish — but unified, deduplicated, and enriched with the fields you actually need to act on.
Four agencies, one schema
CPSC, FDA, NHTSA, and USDA each publish recalls in their own format. We normalize them into a single consistent shape with stable IDs, so you only integrate once.
AI-enriched fields
Severity scoring, affected demographic groups, injury types, hazard summaries, and category tags — generated and reviewed for every recall, so you don't have to build a classification pipeline.
Deduplication built in
When the same product is recalled across multiple lots, agencies, or amendments, we cluster the records and surface the primary variant. No double-counting in your downstream systems.
Historical depth
Every recall we've ever ingested stays queryable with a stable ID and slug history. Compliance and legal teams can audit product portfolios back through years of activity.
{
"id": "ckm8qkp4q0000abcdef123456",
"agency": "CPSC",
"externalId": "25-217",
"title": "TEMI Go Basketball Toys Recalled Due to Choking Hazard",
"url": "https://www.cpsc.gov/Recalls/2025/...",
"recallDate": "2025-08-07",
"status": "ACTIVE",
"hazardLevel": "HIGH",
"primaryCategory": "Toys & Children's Products",
"categories": ["Toys & Children's Products"],
"brands": ["TEMI Go Basketball Toys", "TEMI"],
"models": [],
"upcs": [],
"retailers": ["Amazon"],
"states": [],
"incidentCount": 0,
"injuryCount": 0,
"deathCount": 0,
"remedy": "Stop using and contact TEMI for a full refund.",
"aiSeverityScore": 8,
"aiRiskLevel": "HIGH",
"aiAffectedGroups": ["Infants", "Toddlers"],
"aiInjuryTypes": ["Choking", "Asphyxiation"]
}Trimmed for readability. Full response includes hazard descriptions, source links, image URLs, AI commentary, and FAQ content.
Built for
Teams that need to know about a recall before it becomes a customer complaint, a chargeback, or a headline.
Marketplaces & resale
Match incoming listings against active recalls so recalled products never go live on your platform. Regulators are watching this space — get ahead of it.
Retailer compliance
Continuously reconcile your product catalog against recalls across all four agencies. Replace the manual spreadsheet your compliance team is maintaining today.
Returns & refunds
Programmatically validate recall-driven return requests and route them to the correct workflow. Stop paying agents to Google recall details one at a time.
Insurance & risk
Score product portfolios for recall exposure. Historical data with severity and injury counts feeds straight into your underwriting and claims models.
Brand protection
Manufacturers monitor their own recall footprint and competitors'. Get notified the moment a competitor enters a recall action with the same category or hazard type.
Agents & LLMs
Ground AI safety assistants in authoritative, normalized recall data. Cleaner inputs than crawling four government sites — and citations your users can trust.
Integrate the way that fits
REST API
Query, filter, and paginate recalls by agency, category, brand, hazard level, date range. Standard bearer-token auth, JSON responses, predictable cursor pagination.
Webhooks
Subscribe a URL to recall events filtered by category or brand. HMAC-signed payloads, exponential-backoff retries, idempotency keys. We deliver — you react.
Bulk export
For one-time audits, model training, or warehouse loads. CSV or Parquet snapshots of the full historical corpus, delivered on request.
Request early access
We're onboarding a small group of design partners. Tell us what you're building and we'll set up a 30-minute call to see if there's a fit.
Common questions
How is this different from going to CPSC.gov or openFDA directly?
The agencies publish in four different schemas, with different update cadences, different fields, and no cross-agency deduplication. Building production integrations against all four is several weeks of engineering plus permanent maintenance. We do that work once, so you don't.
Is the data licensed for commercial use?
The underlying recall data comes from US federal agencies and is public domain (17 U.S.C. § 105). Our normalization and AI enrichment is original work licensed to you under the API terms. We'll provide ToS on request.
How fresh is the data?
Ingestion runs daily across all four sources, and we're working with design partners on a near-real-time feed for webhook subscribers. New recalls typically appear in the API within hours of publication by the source agency.
What does it cost?
Design-partner pricing is set per-engagement based on volume and use case. General availability pricing will be tiered and usage-based. Tell us what you'd build and the volumes involved — we'll come back with a number.
Is there an SLA for safety-critical use?
Yes — but with explicit limits. We provide best-available data from authoritative sources, but the API is not a substitute for primary verification. Customers using recall data in safety workflows are expected to maintain their own verification step against the source agency. The full disclaimer is in our ToS.
Who is building this?
RecallRadar is built by Runestone Labs. The recall data platform has been running in production since early 2025, serving the consumer site at recallsradar.com. The API productizes the same data pipeline.