How is this different from hand-rolling API calls?

Each judgment shape gets its own subcommand with documented options and stable JSON field names, plus dry-run previews, pluckable output formats, and malformed responses reported as errors rather than silent passes.

How do I use it as a CI gate?

Match conditions exit 2 while errors exit 1, so chained commands, hooks, and pipelines can distinguish a contradicted claim or failed screen from a configuration problem.

Where does my API key live?

Login stores it in the OS keychain instead of shell files, config shows the effective key source, and per-command docs cover model selection and timeout settings.

Can agents use it during compaction?

Yes. The bundled plugin teaches the agent when to call it and replaces the default compaction summary with the transcript-preserving compact procedure.

Which commands does the jev CLI provide?

verify, screen, classify, extract, find, rerank, match, route, ask, compact, and batch, plus auth, config, models, and update. batch runs any of the others over many rows with a concurrency pool.

What does the dry-run flag do?

It shows exactly what would be sent without calling the API. Add --json to any command for machine-readable output, or --md for PR comments.

What do the jev exit codes mean?

0 is ok, 1 is a usage, config, input, or network error on stderr, and 2 means a --fail-on condition matched. Code 2 is what makes jev usable as a gate in &&, hooks, and CI.

CommunityJev Agents & MCP

Jev CLI — Nasrallah-AL

Command-line toolkit that wraps Jev judgments into scriptable verify, screen, classify, extract, find, rerank, match, route, ask, compact, and batch commands.

See highlights

Open github.com

Stars: 22
Language: TypeScript
License: MIT
Published by: Community
Author: Nasrallah-AL
Last verified: Oct 2026

Nasrallah-AL/jev-cli

User repository on GitHub

View on GitHub

Command-line tool for TypeSafe's Jev AI model

Forks: 3
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

This is one of several Jev listings for the same project. The canonical entry is Jev CLI — shaharia-lab.

About this project

Fast, typed AI judgments from the command line, powered by TypeSafe's Jev model. Jev does not write text. You give it content and a question with a fixed set of answers, and it returns a probability for each answer in a few hundred milliseconds for a fraction of a cent. jev wraps that into commands you can pipe into, script around, and gate CI on. Installed from npm as jevctl. The command is jev. Website: jevcli.vectorz.app.

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

Tagscommunity jevbest catalog-import awesomejev

Works forJev for Classification Jev for Tool calling

Used inJev for Software & DevTools

Languages: TypeScript

Highlights

Installs as jevctl with a jev command, keychain-based login, config inspection, and no-install execution support.
Every command shares input, output, threshold, and timeout conventions with table output by default and JSON for scripts.
Exit codes separate success, usage or network errors, and matched fail-on conditions for gating in hooks and CI.
Ships a Claude Code plugin with slash commands and a compaction hook built on verbatim transcript procedure.

Watch out

Threshold defaults follow TypeSafe cookbooks, so sample your own data with JSON output before tuning for production. 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

JevLoop

Runnable agent harness that moves pick, score and yes-no loop choices to Jev or Laya in milliseconds while reserving the language model for writing, compiling DECISION.md into executable questions.