For developers & AI assistants

Carrier Data API & MCP Server

ProHRHQ publishes FMCSA-derived motor carrier safety data, the same Hoffman Report content on every carrier page, as a free, public API and MCP server. No login, no API key, no paywall. This is the open alternative to gated carrier-safety lookups: anyone (a developer, a broker's internal tool, or an AI assistant like Claude or ChatGPT) can query it directly.

Connect as an MCP server

If you're using Claude, ChatGPT, or any other Model Context Protocol compatible client, add this as a custom connector:

Server URL:  https://prohrhq.com/mcp
Transport:   Streamable HTTP
Auth:        None required
Protocol:    MCP 2025-06-18

In Claude: Settings → Connectors → Add custom connector → paste the URL above. In ChatGPT: enable Developer Mode, add a custom connector with the same URL. Both work today without waiting on any directory listing.

Tools

search_carriers

Search by carrier name, USDOT number, state, or safety grade.

Argument	Type	Required	Notes
`q`	string	no	Carrier name or USDOT number, partial match
`state`	string	no	Two-letter state code, e.g. `TX`
`grade`	string	no	A+, A, B, C, D, or F
`limit`	integer	no	Default 10, max 25

get_carrier_report

Full Hoffman Report for one carrier: safety grade with explanation, BASIC-by-BASIC scores, equipment and cargo profile, and a plain-language verdict. Works for any FMCSA-registered carrier, not only the carriers with a published /carriers/<slug> page.

Argument	Type	Required	Notes
`dot_number`	integer	yes	USDOT number

Example prompts

Once connected, these work as plain conversation with your AI assistant:

"Search ProHRHQ for A-graded carriers in Texas with reefer equipment."
"Look up USDOT 2337430 and tell me if this carrier's safety record is solid."
"Compare the Hoffman Report grades for these three carrier names: [...]."

REST endpoint (no MCP client needed)

For direct HTTP use without an AI client:

GET https://prohrhq.com/api/carriers/find?q=atlantic&state=FL&grade=A%2B&limit=10

Returns JSON: {"hits": [...], "total": N, "error": null}. Same data, same free/no-auth posture.

Rate limits

The MCP tools/call path is rate-limited to protect the server: roughly 20 calls per 60 seconds per client, plus a global ceiling across all clients. initialize and tools/list are unlimited. Limits may tighten if needed: this is a small operation, not an enterprise API, and we ask that integrations be reasonable about call volume.

Data source & accuracy

All carrier data originates from FMCSA's public census, safety, and inspection records (SAFER, SMS). Grading and narrative analysis are ProHRHQ's own, layered on top of the federal record and clearly identified as such. See how FMCSA SMS scoring works for the underlying methodology.

Privacy

No personal data is collected by these endpoints: no login, no tracking, no stored identifiers. See the privacy policy for full detail, including how rate-limit IP handling works.

Questions

Something broken, or want a tool this server doesn't have yet? Contact us.