TypeSafe · Text
Jev 1.13 API pricing
typesafe/jev-1.13
Structured decisions. Jev 1.13 by TypeSafe, routed through INFRO on a TypeSafe account you connect — alongside every other provider you run. TypeSafe bills you for it directly, on that account and at your own rate. INFRO adds routing, failover, request logs, cost attribution and policy around it, and takes no part of what you pay TypeSafe.
Pricing
TypeSafe rate
$0.042
per 1M input tokens, output free, billed to you by TypeSafe
INFRO adds
$0
0% markup on model spend
Per-token rates
$0.042 / 1M input tokens·$0.00 / 1M output tokens
Output tokens are free, so the headline figure above is the input rate, as TypeSafe publishes it.
Calling Jev 1.13
POST /v1/systemone# TYPESAFE_BASE_URL=https://api.infro.io
# TYPESAFE_API_KEY=<your INFRO gateway key>
# TYPESAFE_DEFAULT_MODEL=jev-1.13.0
from typesafe_sdk import Choice, TypeSafeClient
with TypeSafeClient() as client:
response = client.system_one(
state={"document": "I was charged twice. Please fix this."},
questions={
"category": Choice(
instructions="What is this ticket about?",
criteria={"billing": None, "technical": None, "other": None},
),
},
)What Jev answers
- Typed questions about a state: noul (yes or no), choice (up to 255 options), score (2 to 10 levels).
- Probabilities on every answer, and a confidence for choice and score. No generated text.
- Up to 1,000 questions per request through INFRO; TypeSafe's 64k-token request limit applies.
TypeSafe's request body and model names, on your own TypeSafe connection: your INFRO gateway key, spend limits checked before the call, and an estimated cost on every response.
Run Jev through INFRO
What changes in your code, what INFRO checks before TypeSafe is called, and what each response carries.