// The only MCP-native vehicle data API

Vehicle data
that thinks
like your agent.

REST API + native MCP server. Global specs, VIN decoding, crash test data and NHTSA recalls — one bearer token, zero custom wrappers.

Read the docs
Connected to: Carstat MCP live
// Agent-ready Skill

Point your AI Agent at Carstat API.

Use our SKILL.md to give coding agents the exact API context, authentication flow, endpoints, and MCP setup.

382+ makes
3.2K+ models
56K+ configurations
323K+ recalls
237K+ safety ratings
Global Specifications US · EU · Asia · LATAM
// Native MCP Server

Your AI agent,
now knows cars.

Carstat ships a production-ready Model Context Protocol server alongside the REST API. Point VSCode, Cursor, Windsurf, or any MCP-compatible client at your token and it instantly gains 11 structured tools.

available tools
search_vehicles Search & rank vehicles by metric
get_recalls NHTSA recall data
search_recalls Search recalls by make + keyword
get_safety_ratings NHTSA safety ratings by vehicle

+ 7 more tools available

mcp.json
{
  "servers": {
    "carstat-api": {
      "url": "https://api.carstat.app/mcp",
      "type": "http",
      "headers": {
        "Authorization": "Bearer <your-token>"
      }
    }
  }
}

// That's it. Your agent now has:
// · 250+ engineering spec fields
// · VIN → exact spec candidates
// · NHTSA recall lookups
// · NHTSA crash test data
// · Global make/model/trim browse
Works with
VSCode Cursor Windsurf ChatGPT Any MCP client
Full MCP documentation
// How it stacks up

Built for where the
industry is going.

Most vehicle data APIs were designed before AI agents existed. We weren't.

Carstat API you are here
Most vehicle APIs
Native MCP server for AI agents
Global specifications coverage US, EU, Asia & more US-focused
Search by VIN → exact spec candidates
NHTSA recall data on-demand lookup varies
NHTSA safety ratings crash test data varies
EV & battery specs range, charging, connectors limited
Engineering depth suspension, brakes, towing… 250+ fields ~50 fields
REST + MCP in one token no separate auth
// What people build

One API.
Dozens of applications.

AI agents & chatbots

Give any LLM live factory-spec data via MCP. No prompt engineering needed — just structured tools your agent can call directly.

decode_vin("WBA3A5C58DF285669") get_recalls(2022, "Ford", "Bronco Sport")

Automotive apps

Build vehicle configurators, comparison tools, and spec sheets with engineering-grade accuracy. Drill-down to exact trim-level data.

GET /v1/specs/makes/:year GET /v1/specs/:year/:make/:model/:engine

Fleet, insurance & repair

Verify what's actually in a vehicle. Cross-reference VINs against factory specs, decode trim differences, and flag open safety recalls automatically.

GET /v1/specs/vin/:vin GET /v1/recalls/:year/:make/:model

Safety & compliance tools

Build buyer-safety dashboards or compliance workflows using NHTSA 5-star crash test ratings. Surface frontal, side, pole, and rollover scores across 80+ test metrics — down to the trim level.

GET /v1/safety-ratings/:year/:make GET /v1/safety-ratings/:year/:make/:model

Used car marketplaces

Enrich every listing with recall status and crash test scores. Let buyers instantly see whether a vehicle has open NHTSA campaigns or how it performed in safety tests — no extra research required.

GET /v1/recalls/:year/:make/:model GET /v1/safety-ratings/:year/:make/:model

Recall monitoring & alerts

Integrate NHTSA recall data into your product or support workflows. Query by make/model/year, parse defect descriptions, and trigger automated alerts before customers are affected.

GET /v1/recalls/:year/:make GET /v1/recalls/:year/:make/:model
// Explore our data

Three data verticals.
One API.

recalls, safety ratings & specs updated about 9 hours ago
// The data

Not just a VIN lookup.
Engineering-grade specs.

Every configuration returns 250+ fields covering dimensions, drivetrain, EV systems, charging connectors, suspension, towing, and more.

Sample: 2023 Hyundai Kona Electric

Motor & Performance

System Power218 HP
System Torque255 Nm / 188.08 lb-ft
0–100 km/h7.8 sec
Top Speed172 km/h / 106.9 mph
Power / weight123 HP/t · 8.1 kg/HP
Torque / weight143.8 Nm/t · 7 kg/Nm

Dimensions

Length4355 mm / 171.46 in
Width1825 mm / 71.85 in
Height1575 mm / 62.01 in
Wheelbase2660 mm / 104.72 in
Turning circle10.6 m / 34.78 ft
Drag (Cd)0.27

EV Range

WLTP range514 km / 319 mi
Consumption (WLTP)16.8 kWh/100km · 27.04 kWh/100mi
Efficiency6 km/kWh · 3.7 mi/kWh
MPGe125

Battery

Capacity (gross)65.4 kWh
TechnologyLi-Ion
Voltage400 V
LocationBelow Floor

Drivetrain

DriveFWD
Transmission1-speed Auto
ArchitectureBEV
Motor typeSynchronous
Motor locationFront, Transverse
Motor power218 HP / 255 Nm
Layout
Front — Electric motor

Charging

EU 1 port
Front Central
Type 2 AC Fast CCS 2 DC Fast
NA 1 port
Front Central
Type 1 AC Slow CCS 1 DC Fast

Safety & Brakes

Front brakes Vent. Discs
Rear brakes Disc
Assist systemsABS, BAS, EBA, EBD, ESC, HSA, TCS

Suspension

Front McPherson · Stabilizer
Rear Multi-link · Stabilizer
SteeringElectric R&P
Wheel rims17" / 19"

Towing & Load

Curb weight1870 kg
Gross weight2220 kg
Max load447 kg
Roof load100 kg
Towbar nose load100 kg
Trailer (braked 12%)750 kg
Trailer (braked 8%)750 kg
Trailer (unbraked)300 kg
Trunk (min)466 L / 16.46 cu ft
// Discoverable by design

One pattern.
Three data verticals.

Every dataset follows the same browsable flow. Start with available years, drill into makes, then pick a model — works the same whether you're after specs, recalls, or safety ratings.

Vehicle Specs
1
GET /v1/specs/years List available years
2
GET /v1/specs/makes/:year Makes for a year
3
GET /v1/specs/models/:year/:make Models for year + make
4
GET /v1/specs/:modification_id Full 250+ field spec sheet
Recalls
1
GET /v1/recalls/years List available years
2
GET /v1/recalls/makes/:year Makes for a year
3
GET /v1/recalls/models/:year/:make Models for year + make
4
GET /v1/recalls/:year/:make/:model All NHTSA recall campaigns
Safety Ratings
1
GET /v1/safety-ratings/years List available years
2
GET /v1/safety-ratings/makes/:year Makes for a year
3
GET /v1/safety-ratings/models/:year/:make Models for year + make
4
GET /v1/safety-ratings/:year/:make/:model NHTSA crash test data
// Get started

Your first API calls are free.

Free trial available. Access specs, VIN decoding, and recalls from day one. Upgrade when you're ready to ship.

See pricing
// FAQ

Common questions.

What is Carstat API?

Carstat API provides production-grade vehicle specification data, VIN decoding, and NHTSA recall lookups via REST API and a native MCP server. One bearer token gives you access to global make/model/trim data, 250+ engineering spec fields per vehicle, and NHTSA recall data.

Does Carstat support VIN decoding?

Yes. The VIN decoder API accepts any 17-character VIN and returns matching vehicle specifications including year, make, model, engine configuration, and a modification_id you can use to fetch the full 250+ field spec sheet.

Can I access recall data via Carstat API?

Yes. Carstat API proxies NHTSA recall data via /v1/recalls/:year/:make/:model. Results include recall campaign numbers, component descriptions, defect summaries, and remedies — sourced from the NHTSA database.

What is the Model Context Protocol (MCP)?

MCP is an open standard by Anthropic that lets AI agents discover and call structured tools. Carstat ships a native MCP server, so AI assistants like VSCode, Cursor, Windsurf, and ChatGPT can query vehicle data without any custom integration code.

Is there a free trial?

Yes. Carstat API has a free trial. You get immediate API access to integrate and test all the features before committing to a paid plan.

What programming languages are supported?

Any language that can make HTTP requests — Python, JavaScript/TypeScript, Ruby, Go, Java, PHP, and more. Authentication is a standard Bearer token. Interactive API docs with code examples are available.

How many vehicles does Carstat API cover?

Carstat API covers thousands of makes, models, and engine configurations spanning multiple decades of global vehicle production, with 250+ engineering specification fields per variant.