VIN MCP
Provides comprehensive vehicle reports by aggregating data from multiple public sources to decode VINs, check recalls, and view safety ratings. It enables users to validate VINs locally and retrieve technical specifications, fuel economy, and vehicle photos without requiring API keys.
README
VIN MCP
Free VIN decoder for humans and AI. No API keys. No accounts. No limits.
mcp.vin -- Try it now
What It Does
Enter a 17-character VIN and get a comprehensive vehicle report by aggregating six free data sources into a single response:
| Source | Data Provided |
|---|---|
| NHTSA vPIC | Make, model, year, trim, body class, engine specs, transmission, drive type, weight, plant info, 140+ decoded fields |
| NHTSA Recalls | Open recalls with campaign number, component, summary, consequence, and remedy |
| NHTSA Complaints | Consumer complaints with crash, fire, injury, and death statistics |
| NHTSA Safety Ratings | NCAP star ratings for overall, frontal, side, and rollover crash tests |
| EPA Fuel Economy | City/highway/combined MPG, annual fuel cost, CO2 emissions, EV range and charge time |
| IMAGIN.studio | Stock vehicle photos from multiple angles |
Additionally, VIN validation (checksum, WMI country/manufacturer decode, model year) is computed locally with zero API calls.
Quick Start -- Connect via MCP
Smithery (one command)
npx -y @smithery/cli install @keptlive/vin-mcp --client claude
Claude Code (stdio)
Add to your .mcp.json:
{
"mcpServers": {
"vin": {
"command": "npx",
"args": ["-y", "vin-mcp"]
}
}
}
Or if you have the repo cloned locally:
{
"mcpServers": {
"vin": {
"command": "node",
"args": ["/path/to/vin-mcp/server.mjs"]
}
}
}
Claude Desktop / claude.ai (HTTP)
Use the hosted MCP endpoint:
https://mcp.vin/mcp
Add to your Claude Desktop config:
{
"mcpServers": {
"vin": {
"url": "https://mcp.vin/mcp"
}
}
}
MCP Tools
| Tool | Description | Input |
|---|---|---|
decode_vin |
Full VIN decode with specs, recalls, complaints, safety ratings, fuel economy, and photos | { vin: string } |
validate_vin |
Quick local validation -- checksum, WMI country/manufacturer, model year. No external API calls | { vin: string } |
lookup_recalls |
Look up recalls by VIN or by make/model/year | { vin?: string, make?: string, model?: string, year?: number } |
batch_decode |
Decode up to 50 VINs in a single request via NHTSA batch API | { vins: string[] } |
REST API
All endpoints are available at https://mcp.vin or on your self-hosted instance.
| Method | Endpoint | Description |
|---|---|---|
GET |
/api/vin/:vin |
Full decode -- all 6 sources aggregated |
GET |
/api/vin/:vin/validate |
Quick checksum and format validation |
GET |
/api/vin/:vin/recalls |
Recall data only |
GET |
/api/vin/:vin/complaints |
Consumer complaints only |
GET |
/api/vin/:vin/safety |
NCAP safety ratings only |
GET |
/api/vin/:vin/fuel |
EPA fuel economy only |
GET |
/api/vin/:vin/photo |
Redirects to vehicle photo URL |
POST |
/api/batch |
Batch decode (body: { "vins": ["VIN1", "VIN2", ...] }, max 50) |
Rate limits: 30 requests/minute per IP for most endpoints, 60/minute for validation and photos, 5/minute for batch.
Example:
curl https://mcp.vin/api/vin/1HGCM82633A004352
Direct VIN URLs also work -- visit https://mcp.vin/1HGCM82633A004352 to see the web report.
Self-Hosting
git clone https://github.com/keptlive/vin-mcp.git
cd vin-mcp
npm install
node server.mjs --http --port 3200
The server starts in HTTP mode with:
- Web frontend at
http://localhost:3200 - REST API at
http://localhost:3200/api/vin/{vin} - MCP endpoint at
http://localhost:3200/mcp
For stdio mode (Claude Code integration without a web server):
node server.mjs
How VINs Work
A VIN (Vehicle Identification Number) is a 17-character code assigned to every vehicle manufactured since 1981. Each position encodes specific information:
1 H G C M 8 2 6 3 3 A 0 0 4 3 5 2
|_____| |___| | | | |_| |_________|
WMI VDS | | | |yr Sequential
| | | plant
| | check digit
| vehicle attributes
manufacturer ID
- Positions 1-3 (WMI): World Manufacturer Identifier -- country and manufacturer
- Positions 4-8 (VDS): Vehicle Descriptor Section -- model, body, engine, transmission
- Position 9: Check digit -- validates the VIN using a weighted algorithm
- Position 10: Model year code (A-Y, 1-9 on a 30-year cycle)
- Position 11: Assembly plant
- Positions 12-17: Sequential production number
The letters I, O, and Q are never used in VINs to avoid confusion with 1, 0, and 9.
Tech Stack
- Runtime: Node.js 18+
- Server: Express 5
- MCP SDK:
@modelcontextprotocol/sdkwith stdio and Streamable HTTP transports - Frontend: Vanilla HTML, CSS, and JavaScript
- Caching: In-memory LRU with TTL (1h for decodes, 6h for recalls, 24h for ratings and fuel data)
- External dependencies: Zero API keys required -- all data sources are free public APIs
License
MIT
Built for mcp.vin
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。