MCP Demo: VIN Equipment Decoding in Your AI Assistant

Example prompt

"Dekodiere die VIN JMZKF6WMA90393038 und liste die komplette Werksausstattung auf Deutsch auf."

Expected output

The assistant calls decode_vin with lang: "de" and returns structured data — note that every equipment line carries both the translated description and the original wording:

{
  "vin": "JMZKF6WMA90393038",
  "make": "Mazda",
  "model": "Mazda3",
  "year": 2024,
  "equipment": [
    {
      "code": "K6",
      "description": "Sitzheizung für Fahrer und Beifahrer",
      "description_original": "Heated driver and passenger seats"
    },
    {
      "code": "P1",
      "description": "LED-Fernlicht mit automatischem Abblenden",
      "description_original": "LED high beam with auto dimming"
    },
    {
      "code": "R8",
      "description": "Apple CarPlay und Android Auto (kabellos)",
      "description_original": "Wireless Apple CarPlay and Android Auto"
    }
  ]
}

How language is chosen automatically

1. You don't set any language yourself — the assistant reads the language of your conversation and passes it as the lang parameter (write in German → de, write in Danish → da, and so on).
2. Supported codes: de, fr, es, it, nl, pl, pt, sv, da, no, fi, cs. English needs no code — equipment is already in English by default.
3. Translations come from a cached translation layer: phrases already translated are served instantly, new phrases are translated on the fly and stored for next time.
4. The raw, untranslated text is always preserved in description_original, so nothing is ever lost — and any phrase not yet translated simply falls back to the original language.
5. Want to force a specific language? Just say so, e.g. "...auf Polnisch", and the assistant overrides the default.