What is the Model Context Protocol (MCP)?

MCP is an open standard introduced by Anthropic that lets AI assistants connect to external tools and data sources through a uniform interface. Every major MCP-compatible client — Claude Desktop, Claude Code, Cursor, ChatGPT, Continue, Zed, and others — can connect to any MCP server, so a tool you write once is usable everywhere.

What does the try.vin MCP server let an AI assistant do?

The try.vin VIN Decoder MCP server gives any MCP client access to U.S. vehicle data — decode any 17-character VIN into year/make/model/engine via NHTSA's vPIC dataset, and look up NHTSA safety recall campaigns for any 1950-or-later make/model/year. The tools work in both directions: the AI calls them autonomously, and you can also instruct the AI to call them.

Is the try.vin MCP server free?

Yes — the VIN Decoder MCP server is completely free and requires no authentication. It is rate-limited to 120 requests per minute per IP address. Claude's connectors, which reach the server from shared IP addresses, are limited to 120 per MCP session instead, with all sessions from one address held to 1,200 requests per minute together. Every MCP message counts toward these limits. VIN decoding (not recall lookup) is also available through the REST API at /api-docs, which needs a free account and an API token and is limited to 30 requests per minute per token.

Which AI clients can connect to this MCP server?

Any MCP-compatible client that supports HTTP/Streamable transport: Claude and Claude Desktop (as a custom connector), Claude Code, Cursor, VS Code, ChatGPT (as a custom MCP server), Continue, Zed, and any custom Anthropic SDK or OpenAI Responses API integration. The detail page for each server has one-click install links and copy-paste setup for the most common clients.

How is an MCP server different from a REST API?

A REST API is for software you write yourself — you implement the HTTP calls. An MCP server is a contract specifically designed for AI assistants — it advertises its tools, input schemas, and output schemas in a way the AI can discover and call autonomously. Both are built on NHTSA data; today the REST API covers VIN decoding only, while the MCP server also looks up recalls.

Are more MCP servers planned?

Yes. The VIN Decoder is the first server. Coming next are servers for U.S. lemon-law guidance and DMV office lookup.

What data does the VIN decoder return?

The decode_vin tool returns model year, manufacturer (make), model name, trim, series, engine displacement in liters, engine model code, and country of vehicle assembly — sourced directly from NHTSA. Fields are null when NHTSA does not report a value (this is normal for older or low-volume vehicles), and a VIN NHTSA flags but still identifies, such as one with a mistyped check digit, comes back marked partial with NHTSA's warnings.

Model Context Protocol

MCP servers for cars

Free Model Context Protocol servers from try.vin that give Claude, Cursor, ChatGPT, and any other MCP-compatible AI assistant access to U.S. vehicle data — VIN decoding and NHTSA recall lookups, with zero glue code.

One protocol, every client

Works in Claude Desktop, Claude Code, Cursor, ChatGPT, Continue, Zed — anywhere MCP runs.

Free & rate-limited

No API key, no signup — rate-limited to 120 requests per minute per IP address, or per MCP session from Claude's connectors. VIN decoding is also available through the authenticated REST API.

Read-only & agent-safe

Tools are pure lookups — no writes, no side effects. Safe for autonomous agents triaging used-car listings at scale.

Available servers

1 server live today, more on the way.

try.vin VIN Decoder

Live

Decode any 17-character VIN and look up NHTSA recalls — straight from chat.

A free Model Context Protocol server that gives any MCP-compatible AI client (Claude, Claude Code, Cursor, VS Code, ChatGPT, Continue) access to U.S. vehicle data. Three tools: decode_vin (NHTSA vPIC), lookup_recalls (NHTSA campaigns by make, model and year) and lookup_recall_campaign (one NHTSA campaign by its number).

Tools: 3
Transport: HTTP (Streamable)
Auth: None — public, rate-limited

vehiclesvinrecallsnhtsacarsauto

https://try.vin/mcp/vin

Install & docs

What is MCP?

The Model Context Protocol is an open standard from Anthropic that lets AI assistants connect to external tools and data through a uniform interface. An MCP server advertises its tools, input schemas, and output schemas. Any MCP-compatible client — Claude Desktop, Claude Code, Cursor, ChatGPT, Continue, Zed — can discover and call them autonomously, with no glue code on your side.

try.vin runs MCP servers so you can ask any AI assistant to decode a VIN or look up NHTSA recall campaigns without leaving the chat. It's the same data that powers the free VIN check and recall lookup on this site — repackaged for AI-native workflows.

Three steps to install

1
Pick a server
Open a server detail page below to see its tools and copy-paste install snippets for every common AI client.
2
Add it to your client
Use a one-click install link for Claude, Cursor, or VS Code, or paste the server URL into your client. No API key required for the public servers.
3
Restart and test
Restart the client if it asks you to. The new tools appear in the picker — try them with a sample prompt and you're done.

Frequently asked questions

Building software, not chatting?

The same vehicle data is available via an authenticated REST API.

REST API docs

Available servers

try.vin VIN Decoder

What is MCP?

Three steps to install

Pick a server

Add it to your client

Restart and test

Frequently asked questions

01 What is the Model Context Protocol (MCP)?

02 What does the try.vin MCP server let an AI assistant do?

03 Is the try.vin MCP server free?

04 Which AI clients can connect to this MCP server?

05 How is an MCP server different from a REST API?

06 Are more MCP servers planned?

07 What data does the VIN decoder return?

Building software, not chatting?