If you underwrite commercial auto or environmental liability, you already know the pain: Verisk charges five figures for data that the federal government publishes for free.
NHTSA publishes every vehicle recall and safety complaint. The EPA publishes every toxic release from every facility in the country. Both are searchable, structured, and free.
The problem is access. These portals are designed for one-off lookups, not batch risk screening. Pulling 200 VINs through NHTSA's website takes a full afternoon. Cross-referencing facility addresses against EPA TRI records is worse.
What the data actually looks like
NHTSA Vehicle Recalls & Complaints:
- Recall campaigns by manufacturer, component, date
- Consumer complaints with crash/injury/fire indicators
- Investigation status and defect descriptions
EPA Toxic Release Inventory:
- Facility-level chemical release data
- On-site vs off-site disposal quantities
- Carcinogen flags and release trends by year
How to pull it programmatically
Instead of building custom scrapers, you can use pre-built Apify actors that handle pagination, rate limiting, and output formatting:
NHTSA Vehicle Recalls — search by manufacturer, component, or date range. Returns structured JSON with recall descriptions, affected units, and remedy status. $4.00/1K results.
NHTSA Complaints — search by make, model, year, component. Returns crash/fire/injury flags, complaint narratives, and investigation IDs.
EPA TRI Toxic Release — search by facility, chemical, state, or year. Returns release quantities, disposal methods, and carcinogen classifications. $3.50/1K results.
Building a risk screening pipeline
A typical workflow:
- Feed a list of fleet VINs or facility addresses
- Pull NHTSA recall history and open investigations
- Cross-reference facility locations against EPA TRI releases within a radius
- Flag accounts with active recalls, unresolved complaints, or high-volume toxic releases
- Export to your underwriting platform or spreadsheet
Tools like n8n or Make.com can orchestrate this as a recurring workflow — run it weekly against your book of business.
Why this matters for pricing
The data itself is free. The packaging and automation are what save time. Instead of $25K+/year for a Verisk terminal, you can run these queries on demand for pennies per record.
If you're building internal tooling for risk assessment, these APIs plug directly into your existing stack via REST.
I build data automation tools for business research. More at dev.to/avabuildsdata.
Top comments (0)