TypeSafe

Operational· checked every few minutes

Jev 1.14 API

The Jev 1.14 API: TypeSafe's newest decision model, which answers typed questions about your data with calibrated probabilities instead of text, for $0.024 per 1M input tokens with output free.

Price per 1M input tokens43%off

Input: $0.024~~$0.042~~
Output: $0.00

Get your API key Try it in the playground

Model IDjev-1.14

triage.py

# pip install typesafe-sdk
from typesafe_sdk import Choice, Noul, Score, TypeSafeClient
with TypeSafeClient(
    base_url="https://api.zurelay.com",
    api_key="YOUR_ZURELAY_KEY",
    model="jev-1.14",
) as client:
    response = client.system_one(
        state={"ticket": "I was charged twice for October, please refund one of the charges today."},
        questions={
            "refund": Noul(instructions="Is the customer asking for money back?"),
            "team": Choice(
                instructions="Which team should handle this?",
                criteria={"billing": "Charges and refunds", "technical": "Bugs and outages"},
            ),
            "urgency": Score(
                instructions="How urgent is it?",
                criteria=["can wait", "this week", "today"],
            ),
        },
    )
print(response.nouls["refund"].noul)       # 0.98: probability of yes
print(response.choices["team"].choice)     # "billing"
print(response.scores["urgency"].score)    # 1.9: between "this week" and "today"

Base URL https://api.zurelay.com. Works with TypeSafe's Python and TypeScript SDKs. Decisions API docs

Questions: Noul · Choice · Score
Per request: Up to 64K tokens
Output: Typed answers, free
Released: Oct 14, 2026
Status: Operational

Pricing

Jev 1.14 API pricing. A fraction of the list price.

Pay as you go from prepaid credit, with no subscription. Top up from $10. Requests that fail are never billed.

Rate	Zurelay	TypeSafe list	You save
Input per 1M tokens	$0.024	~~$0.042~~	43%off
Output per 1M tokens	$0.00	—	—

Billed on the input tokens the model counts: your state and questions, plus about 270 tokens of its own on every request. Its answers are free, as at TypeSafe. Requests that fail cost nothing.

Savings calculator

Move the sliders to your monthly usage.

Input tokens per month50M

Zurelay

$1.20

TypeSafe list price

$2.10

You save every month$0.90

$10.80 a year

Overview

What is Jev 1.14?

Jev 1.14 is the newest release of Jev, TypeSafe's System One model: you send some content and a set of named questions, and it returns a typed answer for each one with probabilities your code can act on. It is the model TypeSafe's jev-latest alias points to. The Zurelay Jev API serves it on TypeSafe's own endpoint and request shape, so TypeSafe's SDKs work unchanged, at 43% below TypeSafe's list price.

TypeSafe released Jev 1.14 on October 14, 2026, and moved its jev-latest alias, the default in its SDKs, to it. It keeps Jev 1.13's request shape, question types, limits and list price. In our side-by-side tests the two gave the same answers on clear cases, and Jev 1.14 moved its probabilities on borderline ones, so if you've tuned confidence thresholds on Jev 1.13, check them on 1.14 before you switch. Jev 1.13 stays available on Zurelay as jev-1.13.

A request is a state, the content to decide about, as a text, a JSON object or an array, and a map of questions you name. A noul asks a yes or no question and returns the probability of yes, from 0 to 1. A choice picks one option from a set you define, up to 255 options, and returns the probability of each and a confidence. A score places the content on an ordered scale of 2 to 10 levels you describe, and returns a value that can land between levels, the probability of each level and a confidence. Jev reads the state once and answers every question against it in parallel, so ten questions cost little more than one.

Zurelay serves Jev on TypeSafe's own API shape: POST https://api.zurelay.com/v1/systemone with model, state and questions, and the reply carries answers and usage as TypeSafe sends them. TypeSafe's Python and TypeScript SDKs work as they are: set their base URL to https://api.zurelay.com and use your Zurelay key. The model ID is jev-1.14, and jev-latest, jev-1.14.0 and typesafe/jev-1.14 work too. A request can hold up to 64K tokens, the state and every question together, and 32K for the state plus the longest question. TypeSafe's experimental image input, Box questions and long-context header aren't available on Zurelay yet.

Jev is billed on input tokens only, as at TypeSafe: $0.024 per 1M on Zurelay against TypeSafe's $0.042, and its answers are free. The model counts about 270 tokens of its own on every request before your state and questions, so a short question costs well under a thousandth of a cent. Each reply's usage and the x-zurelay-input-tokens header show what it was billed on, every request is in your request log, and requests that fail cost nothing.

Strengths

Where Jev 1.14 shines. And what teams build with it.

Answers your code can use

No parsing free text: every answer is typed, under the name you gave the question, with the probability of each option.

Knows when it's unsure

Calibrated probabilities and a confidence on every choice and score, so you can act on the clear cases and send the rest for review.

Fast and cheap at volume

Usually under a second, billed on input only at $0.024 per 1M tokens: a million short questions cost about $7.

TypeSafe's current release

The model jev-latest points to, with the same request shape as Jev 1.13, so moving up is a one-word change.

Use cases

Ticket and email triage
Route each ticket to a team with a choice, ask whether it's urgent or a refund request with nouls, and rate the customer's frustration with a score, all in one request.
Moderation and risk checks
Flag phishing, spam or policy breaches with a yes/no probability, and use the confidence to decide what a person should look at.
Routing inside agents
Decide which tool, model or workflow handles a request in well under a second, before an expensive model runs at all.
Classifying documents and records
Sort documents, search results or product records into your own categories, or score how relevant each passage is before it reaches a chat model.

Get started

Call Jev 1.14 in three steps.

No waitlist. TypeSafe's own SDKs work as they are: point them at Zurelay and use your Zurelay key.

1
Create a key
Sign up, add credit and create an API key. Set a monthly budget or a rate limit per key if you like.
Get your API key
2
Point your SDK at Zurelay
Set TypeSafe's SDK base URL to this, without /v1: the SDK adds /v1/systemone itself.
https://api.zurelay.com
3
Use jev-1.14
POST a state and named questions to /v1/systemone and get a typed answer with probabilities for each, usually in under a second.
jev-1.14

Works with the tools you already use

Compare

Similar models. All on the same key.

17%off

Jev 1.13

$0.035 per 1M input tokens

Pick Jev 1.13 if you've tuned confidence thresholds on it and aren't ready to retest them.

Jev 1.13 API

90%off

DeepSeek V4.1 Flash

$0.03 / $0.12 per 1M tokens

Pick DeepSeek V4.1 Flash when you need text back: an explanation, a reply or a summary, not a decision.

DeepSeek V4.1 Flash API

61%off

Gemini 3.1 Flash-Lite

$0.10 / $0.59 per 1M tokens

Pick Gemini 3.1 Flash-Lite to classify images, PDFs or audio: Jev reads text only on Zurelay.

Gemini 3.1 Flash-Lite API

FAQ

Jev 1.14 API questions.

Anything else? Ask our team and we’ll answer by email.

How much does the Jev 1.14 API cost?

Zurelay charges $0.024 per 1M input tokens, against TypeSafe's list price of $0.042, so you save 43%. Output is free, as at TypeSafe. Every request counts about 270 tokens of the model's own before your state and questions, so a short request of about 300 tokens costs well under a thousandth of a cent. Requests that fail cost nothing.

What's new in Jev 1.14?

Jev 1.14 is TypeSafe's newest release of Jev, out on October 14, 2026, and the model its jev-latest alias points to. The request shape, question types, limits and price are the same as Jev 1.13's. In our tests it answered clear cases the same way and moved its probabilities on borderline ones. TypeSafe's experimental Box questions, which find something in an image, need Jev 1.14, but images aren't available on Zurelay yet.

Should I use jev-latest or a fixed version?

jev-latest always points to TypeSafe's newest stable release, now Jev 1.14, so its answers can change when a new version ships. If you've tuned confidence thresholds, send jev-1.14 or jev-1.13 instead and move up when you've retested. Each reply's model field names the version that answered.

How do I call Jev 1.14 through Zurelay?

Create a Zurelay API key and POST to https://api.zurelay.com/v1/systemone with model set to jev-1.14, a state and a questions object, each question with a type (noul, choice or score) and instructions, plus criteria for choice and score. With TypeSafe's own SDKs, pip install typesafe-sdk or npm install @typesafe-ai/sdk, then set the base URL to https://api.zurelay.com, without /v1, and use your Zurelay key. The docs at https://zurelay.com/docs/decisions have examples in Python, TypeScript and cURL.

Does TypeSafe's SDK work with Zurelay?

Yes. Zurelay serves Jev on TypeSafe's own endpoint, POST /v1/systemone, with the same request and answer shape. In Python, create TypeSafeClient(base_url="https://api.zurelay.com", api_key="YOUR_ZURELAY_KEY"); in TypeScript, new TypeSafeClient({ baseURL: "https://api.zurelay.com", apiKey }). The SDKs default to jev-latest, which is Jev 1.14. The SDK's models.list call isn't supported: GET /v1/models lists Zurelay's models in OpenAI's format.

How big can a Jev 1.14 request be?

Up to 64K tokens: the state and every question together. The state plus the longest single question must fit in 32K. A choice can have up to 255 options and a score 2 to 10 levels. Zurelay checks a request against these limits before sending it, so a request that's malformed or far too big is refused with a clear error and costs nothing.

How fast is Jev 1.14?

Usually under a second on Zurelay: our tests took 0.4 to 0.7 seconds per request, including one with about 9,400 tokens of logs. Asking several questions in one request is about as fast as asking one, since they're answered in parallel.

Which languages does Jev understand?

TypeSafe trains Jev mainly in English, where it's most accurate. Other languages, CJK scripts included, are handled but not equally well, so test it on your own content and watch the confidence before relying on it for a non-English workload.

Is this TypeSafe's Jev 1.14?

Yes. Your requests run on TypeSafe's Jev 1.14, and Zurelay doesn't swap in a different model or a chat model dressed up as one: before launch we checked its answers and token counts against a second, independent source of TypeSafe's Jev 1.14 release, and they matched. Jev's answers can vary by a point or two between identical calls. Zurelay is an independent service and is not affiliated with TypeSafe.

More models on Zurelay

All models

Stop paying list price.
Start saving today.

Create a key in seconds, point the SDK you already use at Zurelay, and every request costs up to 90% less from the first token.

Get your API key Talk to an engineer

Jev 1.14 API

Jev 1.14 API pricing. A fraction of the list price.

What is Jev 1.14?

Where Jev 1.14 shines. And what teams build with it.

Answers your code can use

Knows when it's unsure

Fast and cheap at volume

TypeSafe's current release

Ticket and email triage

Moderation and risk checks

Routing inside agents

Classifying documents and records

Call Jev 1.14 in three steps.

Create a key

Point your SDK at Zurelay

Use jev-1.14