Why pay for an API call when remainder arithmetic is free?

It does not. It is a joke about API dependence: every valid call costs a request and Jev can misclassify, so reliable free parity checks should use remainder arithmetic instead.

Does passing a huge integer precisely guarantee a right answer?

No. Large integers are accepted precisely as bigint or decimal strings, but precise transport does not guarantee correct classification. Invalid shapes are rejected before any request.

Will it ever return false when the API call itself fails?

Never. There is no cache, local answer, or batch path, and failures reject with typed input, response, or SDK errors. An API failure is never returned as false.

Are the perfect benchmark runs proof it always works?

No. The fixed 20-case run reports correctness, round-trip percentiles, token totals, and per-request evidence, but the digest calls it a small sample with no guarantee for other inputs.

What does a checkOdd call actually return besides the boolean?

A receipt: input, isOdd, parity, confidence, probabilities, model, token usage (input, output and total), durationMs, startedAt, completedAt and an optional requestId. Timing is client-measured and includes network and parsing, not GPU inference time.

Which inputs are rejected locally before any request is made?

Decimals, NaN, infinities, unsafe JS numbers, whitespace, exponent or hex strings and non-numeric strings. Safe integer numbers, bigint and decimal integer strings are accepted, and leading signs or zeroes in integer strings are normalized.

What runtime and credentials does the package need, and how are secrets handled?

Node.js 20.19+ and a Typesafe API key set as TYPESAFE_API_KEY in the server environment. The library does not load .env files itself, ships ESM and CommonJS builds, and warns that the key must not be bundled into browser code.

CommunityJev Benchmarks & Evaluations

Jev-is-odd

Deliberately overengineered TypeScript package that asks Jev whether an integer is odd, with ESM and CommonJS builds, typed answers, and token and timing receipts.

See highlights

Open github.com

Stars: 1
Language: JavaScript
License: MIT
Published by: Community
Author: robipop22
Last verified: Oct 2026

robipop22/Jev-is-odd

User repository on GitHub

View on GitHub

Ask Jev by TypeSafe AI whether a number is odd. TypeScript, real token usage, and latency benchmarks.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

A deliberately overengineered way to ask whether a number is odd. Send an integer to Jev, the decision model from TypeSafe AI. Get a boolean back. Inspired by is-odd-ai. Written in TypeScript, with ESM and CommonJS builds, typed answers, and a receipt for the tokens and time it took. This is a playful AI experiment. Jev can get arithmetic wrong. For reliable, free parity checks, use n % 2 !== 0. This package always asks Jev and returns its decision; it never substitutes local arithmetic.

Quoted from the project's README. Read the full README.

Tagscommunity github

Works forJev for Evaluation

Used inJev for Software & DevTools

Languages: JavaScript

Highlights

Accepts safe integers, bigint, and decimal integer strings; rejects decimals, NaN, infinities, and malformed strings first.
Every valid invocation makes one API call with no cache or local substitute; failures reject rather than returning false.
checkOdd returns parity, confidence, probabilities, token usage, client-measured timing, model, and request ID.
Ships a sequential fixed-case benchmark saving timestamped JSON and Markdown reports plus metrics images for evidence.

Watch out

A playful experiment, not a math utility; Jev can get arithmetic wrong, so use local remainder checks for real parity. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

jev-evals

Rubric-based eval harness that judges LLM and agent outputs with Jev through the Vercel AI Gateway, answering all rubrics for a case in one round trip.

Jev Benchmarks & Evaluations#community#github

Community

NicolasMontone/jev-evalsTypeScript

jev-crash-course

Eleven-level Jev course from hello-world to a ticket-triage capstone, pairing short lessons with runnable scripts against the live API and using LiteLLM so the writing half works with any provider.

Jev Benchmarks & Evaluations#community#github

Community

nadeemcite/jev-crash-coursePython

jev-sim

Wire-compatible Jev re-implementation reading typed decisions from an LLM's next-token logits, plus a reproducible JevBench-based comparison of that approach against Jev on identical items.

Jev Benchmarks & Evaluations#community#github

Community

dashbi1/jev-simPython

apple-silicon-systemone-benchmark

Native Apple Silicon harness comparing local decision models through one decision interface, running models serially and emitting an offline dashboard with latency and accuracy.

Jev Benchmarks & Evaluations#community#github

Community

Giamme/apple-silicon-systemone-benchmarkPython

jev-verify

Node.js CLI that checks whether published Jev outputs were actually produced by Jev: it recomputes the Yurin confidence identity and the expected-score formula over a repository's artifacts and flags fixtures that violate them. No API calls, no dependencies, Node 18+.

Jev Benchmarks & Evaluations#community#calibration#verification

Community