How to get access to Jev and make your first call

1. Get on the early-access list

Jev came out of stealth on September 15, 2026, and access is still being rolled out in batches. TypeSafe is running an early-access waitlist on its official console and issuing API keys to approved developers in waves.

Start at typesafe.ai to join the queue.

At launch there was no publicly documented free-credit or trial program, so budget for paid usage from your first request. Given the price, that is a low bar — see pricing.

2. Or skip the queue via an aggregator

If you are still waiting on a first-party key, Jev is already resold through gateways you may already be using:

Going through a gateway means the gateway's billing, rate limits and privacy terms apply — not TypeSafe's. Worth a read before you route production traffic through one.

3. Make your first request

The endpoint is POST https://api.typesafe.ai/v1/systemone, and the early-access model route is jev-latest. Note this is not the OpenAI chat-completions shape — you send a state plus questions, so you need TypeSafe's client (or a small bespoke one) rather than pointing an OpenAI SDK at a new base URL.

Install

pip install "typesafe-sdk>=0.5.7"

Python

import os
from typesafe_sdk import Choice, Noul, NoulCriteria, Score, TypeSafeClient

client = TypeSafeClient(api_key=os.environ["TYPESAFE_API_KEY"], timeout=120.0)

response = client.system_one(
    model="jev-latest",
    state="The payment failed twice and the customer is asking for a refund.",
    questions={
        "department": Choice(
            instructions="Which team should handle this?",
            criteria={
                "billing":   "Payments, invoicing, refunds",
                "technical": "Bugs, outages, integrations",
                "sales":     "Pricing, upgrades, new accounts",
            },
        ),
        "urgency": Noul(
            instructions="Does this convey urgency?",
            criteria=NoulCriteria(
                true="Explicitly time-sensitive",
                false="No urgency expressed",
            ),
        ),
        "frustration": Score(
            instructions="How frustrated is the customer?",
            criteria=["Calm", "Frustrated", "Very angry"],
        ),
    },
)

answers = response.answers
print(answers["department"].choice)      # -> "billing"
print(answers["department"].confidence)  # -> 0.0-1.0
print(answers["urgency"].noul)           # -> 0.0-1.0
print(answers["frustration"].score)      # -> "Frustrated"
print(response.usage.input_tokens)

JavaScript / TypeScript

An official JavaScript SDK ships alongside Python under the same typesafe-sdk package name. TypeSafe's published examples are Python-first, so for exact JS signatures check the SDK section of the docs rather than transliterating the Python above.

4. Things that will bite you

• It will not write anything. If your use case ends in a sentence, Jev is the wrong half of the pipeline. Pair it with an LLM.
• Options must be known at request time. Choice picks from the set you send, so a dynamic action space means building that list per call.
• Use the confidence value. A winning option at 0.34 confidence is a coin flip with extra steps — gate on a threshold and fall back to an LLM or a human.
• ~32,000 tokens per request per request. Long documents need chunking or a retrieval step in front.
• Text only at launch. Text only — no image or audio input at launch