CarrierScore MCP Server
Provides FMCSA motor-carrier risk intelligence over MCP, with tools for carrier lookup, risk scoring, and timestamped evidence reports for AI agents and freight brokers.
README
<img src="https://carrierscore.io/icon-512.png" alt="CarrierScore" width="96" align="right">
CarrierScore MCP Server
FMCSA motor-carrier risk intelligence for AI agents, over the Model Context Protocol.
CarrierScore answers "how risky is this carrier, and what was knowable about it on a given date?" for freight brokers and AI booking/dispatch agents. It builds longitudinal risk scores (0–100, higher = riskier) and timestamped carrier-selection evidence reports ("Montgomery files") from daily archives of public FMCSA / US DOT safety data: census, inspections, crashes, out-of-service orders, operating authority, and insurance filings.
Website: carrierscore.io · Agent docs: carrierscore.io/llms.txt
Hosted endpoint (recommended)
No install needed — a hosted Streamable HTTP MCP endpoint is live:
https://mcp.carrierscore.io/mcp
Keyless requests get a rate-limited free tier (lookups, scores, text evidence reports). Paid tiers add carrier-list monitoring, structured JSON evidence reports, and higher volume — see pricing. Paid tiers authenticate with an X-API-Key header (or CARRIERSCORE_API_KEY when self-hosting).
Tools
| Tool | What it does |
|---|---|
carrier_lookup |
Carrier identity by US DOT number: legal name, DBA, operating status, FMCSA safety rating, fleet size, address, registration dates. Use first to confirm a DOT is real and matches the rate confirmation. |
carrier_score |
The CarrierScore risk score (0–100, higher = riskier) with full component breakdown, hard flags (active OOS order, no active insurance, reincarnated-carrier link), and data-sufficiency indicator. |
montgomery_file |
Timestamped carrier-selection evidence report — the documented artifact a broker retains at the moment of booking (post Montgomery v. Caribe Transport II). Text format on the free tier; JSON on paid tiers. |
monitor_carriers |
Batch score summary + flags for up to 100 DOT numbers per call — screen candidates for a load or re-check an active roster. |
Connect from a client
Claude Code
claude mcp add --transport http carrierscore https://mcp.carrierscore.io/mcp
Claude Desktop / claude.ai
Settings → Connectors → Add custom connector → URL https://mcp.carrierscore.io/mcp.
Cursor
.cursor/mcp.json:
{
"mcpServers": {
"carrierscore": {
"url": "https://mcp.carrierscore.io/mcp"
}
}
}
Generic MCP client (Streamable HTTP)
{
"mcpServers": {
"carrierscore": {
"type": "streamable-http",
"url": "https://mcp.carrierscore.io/mcp"
}
}
}
Self-hosting
The server in this repo is a thin Streamable HTTP client over the CarrierScore REST API (https://api.carrierscore.io).
npm install
npm run build
CARRIERSCORE_API_URL=https://api.carrierscore.io npm start
# MCP endpoint: POST http://127.0.0.1:8322/mcp
Environment variables:
| Variable | Default | Purpose |
|---|---|---|
CARRIERSCORE_API_URL |
http://127.0.0.1:8321 |
Base URL of the CarrierScore REST API |
CARRIERSCORE_API_KEY |
(unset) | Optional API key for paid tiers; unset = rate-limited free tier |
PORT |
8322 |
Port for the MCP HTTP endpoint (POST /mcp) |
Disclaimer
CarrierScore scores are computed statistical indicators built from public FMCSA data under a documented, transparent methodology (population percentiles plus hard flags). They are not safety fitness determinations, are not endorsed by FMCSA, and do not substitute for a carrier's official safety rating or your own judgment. Every score and evidence report embeds this disclaimer — retain it when storing or quoting results.
License
MIT © SYMBOLIQ LLC. Contact: business@colesprout.com
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。