Intent Classification & Routing API — millisecond decisions

Intent classification is the highest-volume decision in most LLM stacks: every incoming message, ticket or command must be routed to the right handler before any generative model runs. Doing it with a chat model means paying for output tokens and waiting on generation; doing it with a decision model returns a typed choice across up to 255 candidates in a single forward pass.

Why decision models

• Typed choice answers — branch your code on the result directly
• Up to 255 routing candidates per question
• Free output tokens: routing never pays for generation
• One endpoint for the whole catalog — switch models per request

EasyJev serves the whole System One catalog behind one endpoint compatible with the official typesafe-sdk: swap base_url and api_key, pick any model per request, and settle token-precise bills. Try it live in the Playground before writing a line of code.

Runnable example

curl https://api.easyjev.online/v1/systemone \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "typesafe/jev-latest",
    "state": {"user": {"plan": "pro"}},
    "questions": {
      "is_premium": {"type": "noul", "criteria": {"true": "…", "false": "…"}}
    }
  }'