Car Check in ChatGPT
Recalls, ratings and mpg. The MCP server is https://cars.openkrill.app/mcp. No key. The steps below are only for ChatGPT. A call to decode_vin was checked against that server on 2026-10-01.
What this server answers
Check a US car's public record before you buy or while you own it: decode a VIN, see its recalls, crash test ratings and owner complaints, and compare fuel economy.
Ask "what car is VIN 1HGCM82633A004353?", "any recalls on a 2018 Toyota Camry?", "give me a report on a 2016 Honda Civic" or "what mpg does a 2022 Toyota RAV4 hybrid get?".
Car Check reads two free official sources: the National Highway Traffic Safety Administration (NHTSA) for VIN decoding, recalls, 5-star crash test ratings and consumer complaints, and the US Environmental Protection Agency's fueleconomy.gov for mpg, mpge, electric range, tailpipe CO2 and estimated annual fuel cost.
The report answers in one call: the vehicle, its crash ratings, how many recalls NHTSA lists and the newest ones, how many complaints were filed and which components they name most often, and its fuel economy.
Complaints are shown only as a count and the components named. Individual complaint texts and VIN fragments are never returned, and a complaint is a report from an owner, not a confirmed defect.
It covers vehicles sold in the US. Recalls and complaints are listed by year, make and model, not by VIN, so for one specific car confirm at nhtsa.gov/recalls. A VIN you enter is sent to NHTSA to decode it and is never stored.
Car Check is a data lookup. It is not a valuation, a vehicle history or accident report, an inspection, or safety, mechanical or buying advice.
Data from NHTSA (nhtsa.gov) and the US EPA (fueleconomy.gov), which are public US government sources. It stores nothing about you: answers are cached for a day, and the only lasting record is a daily count of calls per tool.
What it can do
- Decodes a 17-character VIN into year, make, model, trim, body style, engine and fuel, with NHTSA's own error codes
- Lists NHTSA recalls for a year, make and model with the component, consequence and remedy
- Shows NHTSA 5-star crash test ratings: overall, front, side and rollover
- Counts owner complaints filed with NHTSA and names the components reported most often
- Shows EPA fuel economy: city, highway and combined mpg, mpge, electric range and annual fuel cost
- Builds a one-call report from a VIN or from year, make and model
- Covers US vehicles only and never gives valuation, history or safety advice
Tools
decode_vin, Decode a VIN. Use this when the user gives a 17-character vehicle identification number (VIN) and wants to know what car it is, such as "what car is 1HGCM82633A004353?". Returns the model year, make, model, trim, body class, engine, fuel, plant country and NHTSA's error codes and error text, which flag a wrong check digit or an incomplete decode. US-market vehicles. A data lookup, not a history report.get_vehicle_report, Get a vehicle report. Use this when the user wants an overview of a US car, such as "give me a report on a 2016 Honda Civic". Pass a VIN, or the year, make and model. Returns NHTSA 5-star crash ratings per variant, the number of recalls and the newest ones, the number of owner complaints with the components named most often, and EPA fuel economy. Complaint texts and VIN fragments are never returned; complaints are not confirmed defects. Not a valuation, history report or advice.list_recalls, List recalls. Use this when the user asks about recalls for a car model, such as "any recalls on a 2018 Toyota Camry?". Pass the year, make and model and optionally a limit of up to 25. Returns each recall's campaign number, date, component, summary, consequence and remedy, and whether NHTSA says to park the car. Recalls are listed by year, make and model, not by VIN: to check one specific car, use nhtsa.gov/recalls. A model with no recalls listed is not a safety guarantee.get_fuel_economy, Get fuel economy. Use this when the user asks about mpg, electric range or fuel cost for a car, such as "what mpg does a 2022 Toyota RAV4 hybrid get?". Pass the year, make and model, and optionally words such as AWD or hybrid to narrow it. Returns the EPA's city, highway and combined mpg (mpge for electric vehicles) for each configuration, with electric range, tailpipe CO2 and estimated annual fuel cost, and lists the configurations not shown. EPA estimates, not a promise of real-world mileage.
Add Car Check in ChatGPT
ChatGPT does not read a local config file. A custom MCP server is added in the product, and only after developer mode is on. These steps were checked against OpenAI's plugin connect guide and the ChatGPT MCP guide on 2026-10-01. Developer mode can be hidden by a workspace policy, so a Team or Enterprise admin may have to allow it first.
- Open ChatGPT settings, then Security and login, and turn on Developer mode.
- Open ChatGPT plugins and select the plus button.
- Enter a name such as "Car Check" and a one-line description of what you want it to answer.
- Under Connection, choose the public endpoint and paste
https://cars.openkrill.app/mcp. Include the/mcppath. Do not add a tunnel id: this server is already public. - Create the connection and review the tools ChatGPT lists. You should see
decode_vin,get_vehicle_report,list_recalls,get_fuel_economy. The feedback tools on the same server are optional and not required for the lookup. - Start a new conversation, add the connection from the tools menu, and ask: Give me a report on a 2016 Honda Civic
No API key and no OAuth client id are required. The server answers without a login. Calls are limited per network and per day, and a limit comes back as a tool error rather than a partial answer. If the tool list is stale after a server change, open the connection and select Refresh, then start a new conversation.
The ChatGPT desktop app, Codex CLI and IDE extension share a different config. In the desktop app, open Settings, then MCP servers, choose Add server, select Streamable HTTP, and paste the same URL. In ~/.codex/config.toml the equivalent is a table named mcp_servers.cars with url = "https://cars.openkrill.app/mcp". ChatGPT on the web does not read that file. A published directory listing is also a different path: this page is the developer-mode connection, not a claim that Car Check is already in the ChatGPT directory.
Treat the server as untrusted until you have read a few answers. A remote MCP server sees the arguments the model sends. Car Check is read-only and documents what it sends onward on the privacy page.
A call checked on 2026-10-01
This is the sample VIN in the plugin's starter prompts. NHTSA decodes it; it is not a history report. The request below was posted to https://cars.openkrill.app/mcp as tools/call. HTTP 200. Source: the Car Check server, read 2026-10-01.
{
"method": "tools/call",
"params": {
"name": "decode_vin",
"arguments": {
"vin": "1HGCM82633A004353"
}
}
}- status. ok
- notice. US vehicles only. This is public vehicle data from NHTSA and the US EPA, not a valuation, vehicle history or condition report, and not safety, mechanical or buying advice. Recalls and complaints are listed by year, make
Repeat the call yourself if you need a newer reading. Cached answers expire. A rate limit is not a result: wait and try again. Nothing in the call is a ranking, a filing, or advice.