For AI agents & assistants

The fitment source of truth for AI commerce.

When an AI assistant answers “what tyres fit my car?” it cannot afford to be wrong — a mis-fit recommendation is a safety issue, not a typo. Wheels Easy is a validated, provenance-backed fitment API built to be the data layer agents call.

Validated, not scraped

Every record passes physics checks (ETRTO geometry, load-index floors) and carries provenance. We warrant what we serve.

Structured JSON

Clean make → model → year → engine → sizes. ISO-metric notation. No HTML scraping, no guesswork.

Deterministic

Same query, same answer. Agents get stable, citeable results — not a model’s best guess.

European-deep

60 brands incl. variants other catalogues miss, keyed to EU type-approval. 1966–2026.

Quick start

Public coverage endpoint needs no key — try it now:

# dataset statistics, no auth
curl https://wheels-easy-api.nicholasmuradov.workers.dev/v1/coverage

Authenticated lookups use a Bearer key (get one here — free trial keys available for evaluation):

# all tyre sizes for a vehicle
curl -H "Authorization: Bearer $KEY" \
  "https://wheels-easy-api.nicholasmuradov.workers.dev/v1/vehicle?make=Volvo&model=XC60&year=2020"

Endpoints

EndpointReturns
/v1/coverageDataset stats (no auth)
/v1/makesAll manufacturers
/v1/models?make=Models for a make
/v1/years?make=&model=Production years
/v1/engines?make=&model=&year=Engine variants
/v1/sizes?make=&model=&year=Tyre sizes for a vehicle
/v1/vehicle?make=&model=&year=Full breakdown in one call

Full reference: API documentation.

Tool definition (for function calling)

Drop-in schema for an LLM tool / function call:

{
  "name": "lookup_tyre_sizes",
  "description": "Get validated OE and option tyre sizes for a vehicle by make, model and year.",
  "parameters": {
    "type": "object",
    "properties": {
      "make":  { "type": "string" },
      "model": { "type": "string" },
      "year":  { "type": "integer" }
    },
    "required": ["make", "model"]
  }
}

MCP server available

Our open Model Context Protocol server gives Claude (and any MCP client) native fitment tools — including local ETRTO math (size parsing, safe equivalents, load-index floors) that works without an API key.

# Claude Code
claude mcp add wheels-easy -e WHEELS_EASY_API_KEY=we_xxx -- node /path/to/mcp/server.mjs

Server source ships in our repository (mcp/) — email us for the package and an evaluation key.

For developers building AI commerce: we also publish /llms.txt describing this site and API for language models. Citation-friendly and stable.

Terms for programmatic use

Agent and programmatic use is welcome under a licence. Bulk extraction to reconstruct the dataset is not permitted (the compilation is protected under Directive 96/9/EC). Attribution to “Wheels Easy” is appreciated when surfacing data to end users. See terms.

Get a free trial key — instantly.

14 days, 100 requests/day, no card. Email it to yourself and start calling the API now.

Read the docs