US Recalls Scraper: FDA, NHTSA and CPSC Recalls in One Dataset (Free)

US Recalls Scraper collects US government product recall information into a unified dataset. It covers FDA food, drug, and medical device recalls, NHTSA vehicle recalls, and CPSC consumer-product recalls.

Choose a collection mode, specify a date range, and optionally filter by keywords, recalling firms, vehicle details, or campaign numbers. The scraper also supports VIN decoding and can retrieve matching vehicle recalls.

Export structured records as JSON, CSV, or Excel for product safety monitoring, compliance research, automotive recall checks, and reporting.

Use Cases

• Retailers and Amazon sellers: Monitor product recalls to support product safety checks.
• Compliance and QA teams: Collect food, drug, and medical device recall records for a selected date range.
• Car dealers and fleets: Look up vehicle recalls by make, model, year, or campaign number.
• VIN recall checks: Decode vehicle identification numbers and retrieve matching recall campaigns.
• Insurers: Research vehicles associated with specific recall campaigns.
• Journalists and researchers: Track recalling firms, product hazards, and recall trends.
• Consumer applications: Incorporate structured vehicle and product recall records into data workflows.

Collection Modes

Mode	Input	Results
all	Date range and optional keyword	FDA food, drug, and device recalls, plus CPSC recalls
fda	Categories, keyword, firm, classification, and dates	FDA enforcement recall records
cpsc	Date range and optional keyword	Consumer-product recall records
vehicle	Make, model, year, or campaign numbers	NHTSA vehicle recall records
vin	One or more 17-character VINs	Vehicle decoding results and optional matching NHTSA recalls

USDA FSIS meat and poultry recalls are not included because the source returns HTTP 403 from a datacenter IP.

Input Parameters

Parameter	Type	Default	Description
mode	string	all	Select all, fda, cpsc, vehicle, or vin.
fdaCategories	array	food, drug, device	FDA datasets to search.
keyword	string	Empty	Filters FDA product descriptions and reasons, or CPSC titles and descriptions.
recallingFirm	string	Empty	Filter FDA recalls by recalling firm.
fdaClassification	string	any	Select FDA Class I, Class II, Class III, or any classification.
dateFrom	string	30 days ago	Start date for recall collection.
dateTo	string	Today (UTC)	End date for recall collection.
vehicles	array	Honda Accord 2020	Vehicle objects containing make, model, and modelYear.
campaignNumbers	array	Empty	NHTSA campaign numbers to look up.
vins	array	Empty	Vehicle identification numbers, one per item.
vinCheckRecalls	boolean	true	Retrieve matching NHTSA recalls for decoded vehicles when possible.
maxItems	integer	200	Maximum number of recall and VIN rows. Hard maximum: 5,000.
includeRaw	boolean	false	Include the original source record in the raw field.
openFdaApiKey	string	Empty	Optional key to increase the available daily request allowance. The key is not saved in dataset records or logs.
timeoutSecs	integer	300	Maximum runtime in seconds.
requestTimeoutSecs	integer	30	Per-request timeout. CPSC requests use at least 90 seconds.

Example Input

{
  "mode": "all",
  "keyword": "listeria",
  "dateFrom": "2026-09-12",
  "dateTo": "2026-10-12",
  "maxItems": 200
}

For vehicle mode, use the default vehicle or provide your own make, model, and model year. You can also supply campaign numbers.

For VIN mode, provide 17-character VINs. A VIN with an invalid check digit may still return decoded information, with errorCode and errorText indicating the result.

What Data Does This Actor Extract?

Each recall record follows a consistent schema. The recallTitle field contains the display title and is also stored as title.

Example Recall Record

{
  "rowKind": "recall",
  "source": "nhtsa",
  "recallId": "20V771000",
  "recallTitle": "HONDA ACCORD 2020 - ELECTRICAL SYSTEM:BODY CONTROL MODULE:SOFTWARE",
  "title": "HONDA ACCORD 2020 - ELECTRICAL SYSTEM:BODY CONTROL MODULE:SOFTWARE",
  "reason": "Honda is recalling certain 2018-2020 Accord Sedan vehicles.",
  "company": "Honda (American Honda Motor Co.)",
  "recallDate": "2020-10-12",
  "remedy": "Honda will notify owners, and dealers will update the BCM software, free of charge.",
  "consequence": "The software can increase the risk of a crash.",
  "vehicle": {
    "make": "HONDA",
    "model": "ACCORD",
    "modelYear": "2020",
    "component": "ELECTRICAL SYSTEM:BODY CONTROL MODULE:SOFTWARE"
  }
}

Output Fields

Field	Type	Description
source	string	Source category: fda_food, fda_drug, fda_device, nhtsa, or cpsc.
recallId	string	FDA recall number, NHTSA campaign number, or CPSC recall number.
recallTitle	string	Short recall title, also stored as title. FDA product text is limited to 120 characters in this field.
productDescription	string	Product description or additional summary text.
reason	string	Recall reason or CPSC hazard description.
hazard	string or array	CPSC hazard information, when available.
classification	string	FDA recall classification.
status	string	FDA recall status, when available.
company	string	Recalling firm, manufacturer, or importer.
companyCountry	string	Company country, when available.
companyState	string	FDA firm state, when available.
recallDate	string	Recall date in ISO format.
reportDate	string	Report date in ISO format.
terminationDate	string	FDA termination date, when available.
distribution	string	FDA distribution information.
quantity	string	Quantity or units, preserving source formatting where applicable.
remedy	string	Repair, replacement, refund, or other remedy information.
consequence	string	Potential vehicle consequence, when available.
injuries	string	CPSC injury information, when available.
url	string	CPSC recall page, when provided.
images	array	CPSC product image references.
vehicle	object	Vehicle make, model, year, component, and unit information.
identifiers	object	Identifiers such as NDC, brand name, UPC, and event ID.
raw	object	Original source record, included when includeRaw is enabled.

VIN Output

VIN records use rowKind: "vin_decode" and may include:

• vin
• make
• model
• modelYear
• trim
• bodyClass
• vehicleType
• manufacturer
• plantCountry
• fuelTypePrimary
• errorCode
• errorText
• isValid
• recallCount

The isValid field is true only when errorCode is 0.

Error Records

Failed inputs are saved as error rows with fields such as sourceInput, error, and statusCode. One invalid VIN does not stop the entire run.

A “No matches found” response from FDA is treated as an empty result rather than a request failure.

The dataset includes separate views for Recalls, Vehicles, VINs, and Errors.

How It Works

1. Read the selected mode, date range, filters, and record limit.
2. Use default dates covering the last 30 days through the current UTC date when dates are not supplied.
3. Retrieve records for the selected source and collection mode.
4. Normalize dates from supported formats, including YYYYMMDD, MM/DD/YYYY, and ISO timestamps.
5. Standardize the recall fields across sources.
6. Remove duplicate recall records while preserving separate vehicle-specific campaign records when appropriate.
7. Sort recall records by date, newest first.
8. Stop when the record limit or runtime limit is reached.
9. Save successful records and applicable error rows to the dataset.

Dataset and Export

The actor produces structured data suitable for:

• JSON, CSV, and Excel exports.
• Product safety and compliance reports.
• Vehicle fleet and VIN recall research.
• Recall history tracking.
• Data analysis with Python and Pandas.
• Scheduled monitoring and downstream automation.

Schedules and Alerts

Schedule recurring runs to monitor a keyword, recalling firm, product category, or vehicle list. Scheduled datasets can be connected to supported notification and automation workflows.

For example, a retailer can schedule regular recall checks and review newly collected records, while a fleet manager can monitor recalls associated with selected vehicle models.

Limits and Data Usage

• FDA enforcement records may be refreshed approximately weekly. Results depend on what has been published by the agency.
• Without an optional FDA key, the actor observes a limit of 1,000 requests per day per IP for FDA requests.
• With an FDA key, the documented daily allowance is higher, subject to the applicable service limits.
• FDA request pacing is at least 0.3 seconds between calls.
• FDA page size is limited to 1,000 records per request, and the skip value does not exceed 25,000.
• Larger result sets may be split into date windows.
• CPSC requests can take longer than other requests, so they use a minimum request timeout of 90 seconds.
• CPSC date ranges are processed in windows of 31 days. The end date and keyword are applied locally.
• VINs are decoded individually.
• The maximum number of rows per run is 5,000.
• The maximum overall runtime is 3,600 seconds.
• Only information returned by the relevant source can be collected; unpublished or unavailable records may not appear.

Frequently Asked Questions

Does this scraper include USDA meat and poultry recalls?

No. USDA FSIS meat and poultry recalls are excluded because the service returns HTTP 403 from a datacenter IP.

Do I need an FDA API key?

No. An FDA key is optional. It increases the available daily request allowance, while runs without a key remain subject to the lower request limit.

How does a VIN recall check work?

The actor decodes the supplied VIN. When the decoded vehicle has a make, model, and year, the scraper can retrieve matching NHTSA recall records if vinCheckRecalls is enabled.

Can a VIN with an invalid check digit return data?

Yes. The decoder may return vehicle information while providing errorCode and errorText to indicate the validation result.

How fresh is the recall data?

Freshness depends on the source. FDA enforcement information may be refreshed approximately weekly, while CPSC and NHTSA results depend on the records available when the run executes.

Can I export the results?

Yes. Export the dataset in JSON, CSV, or Excel format.

Can I receive recall alerts?

Yes. Schedule recurring runs and connect supported integrations to send notifications or transfer new records to other systems.

What happens if one request fails?

The actor records relevant errors in the dataset and continues processing other inputs whenever possible.

Pricing

The actor is free to run. Apify platform usage is billed separately according to the applicable plan.

Integrations

The dataset can be used with supported integrations, spreadsheet tools, automation platforms, databases, and webhooks.

Trademark Notice

The U.S. Food and Drug Administration, the National Highway Traffic Safety Administration, the U.S. Consumer Product Safety Commission, and their respective names, logos, and trademarks belong to their owners. Their use in this documentation is solely to identify the data sources and does not imply affiliation, sponsorship, endorsement, or partnership.

Disclaimer

This actor is an independent tool and is not affiliated with, endorsed by, sponsored by, or officially associated with the U.S. Food and Drug Administration, the National Highway Traffic Safety Administration, or the U.S. Consumer Product Safety Commission.

Users are responsible for complying with applicable source terms, rate limits, licensing conditions, and attribution requirements.

Is Web Scraping Legal?

Collecting information that is already public is generally legal. Scraping is not the same thing as hacking, and there is no single law that bans it. The limits are the usual ones: do not break into accounts, do not take private or confidential data, and be careful with personal data and copyrighted works. A site's terms can also restrict how you access it. Facts such as a public name, price, or rating are not the same thing as republishing someone's article or photo.

This actor collects publicly accessible recall information. It does not log in and it does not bypass access controls.