Why use a decision model instead of prompting a chat LLM?

Jev returns typed values with probabilities instead of text, so there is nothing to parse. Low-confidence categories fall into a separate review list controlled by a threshold flag.

What input format does it expect, and is it one call per message?

One message per line with blanks and hash-prefixed lines ignored. All three questions run in parallel per message, with concurrent workers configurable by flag.

What happens on rate limits or a bad key?

Rate-limit, overload, and server errors retry a few times with a wait; bad keys and other errors fail fast and appear in the failure list. A base-URL variable can point at a proxy or test server.

How do I keep weird messages from misrouting?

Keep an other option in every choice question so odd messages have somewhere to go, then inspect the review list and tune the threshold for your own traffic.

What questions does it ask each message?

Three: a Choice for the kind of message (bug, billing, feature_request, account or other), a Score for urgency from 0 to 3, and a Noul for whether the sender sounds frustrated. All run in one request and are judged separately.

What are the flags and the output?

Flags are -file (messages.txt), -html (report.html; pass empty to skip), -threshold (0.8) and -workers (4). It prints a table sorted by urgency, puts low-confidence categories in a review list, lists failures, and writes a color-coded HTML report with a confidence bar per message.

What do I need to build and run it?

Go 1.21 or newer and a TypeSafe API key, with no third-party Go packages. Set the key in a .env file or the terminal, where the environment wins, then run go run . . TYPESAFE_API_BASE can point it at a proxy or local test server.

CommunityJev Agents & MCP

Jev Triage — boldbug1

Small Go triage tool that sends each message plus three questions to Jev in one request, then prints an urgency-sorted table, a low-confidence review list, and an HTML report.

See highlights

Open github.com

Stars: 3
Language: Go
License: MIT
Published by: Community
Author: boldbug1
Last verified: Oct 2026

boldbug1/jev-triage

User repository on GitHub

View on GitHub

Message triage CLI in Go, built on the Jev decision model from TypeSafe AI. Categorizes messages, scores urgency, and flags low-confidence ones for human review.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

A small command-line tool, written in Go, that reads a list of messages and sorts them by urgency. It uses Jev, a decision model from TypeSafe AI, to answer three questions about each message: It prints a table in the terminal, puts low-confidence messages in a separate "review by hand" list, and writes an HTML report.

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

Tagscommunity awesomejev catalog-import

Works forJev for Classification Jev for Ranking

Used inJev for Software & DevTools

Languages: Go

Highlights

Asks category, urgency score, and frustration probability about each message together.
Sorts the terminal table by urgency with confidence bars and a review section.
Writes the same results to a color-coded HTML report, skippable by flag.
Uses only the standard library with retries and an overridable API base URL.

Watch out

Per-request questions are judged separately and the default threshold is only a starting point for your data. 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.

Other listings of this project

Forks and copies of this project in the directory. This entry is the canonical one.

jev-triage

Probot GitHub app that triages issues and pull requests with two Jev fan-outs for category, impact, completeness, spam, and duplicates, then labels and comments once.

Jev Tools & CLIs#community#github

Community

ashafizullah/jev-triageTypeScript

jev-triage

GitHub Action that labels newly opened issues with Jev, returning a label plus confidence and flagging low-confidence cases for a human or an LLM instead of guessing.

Jev Tools & CLIs#community#github

Community

CMaintz/jev-triage

More Jev resources like this

jev-judge-mcp

MCP server over stdio that gives coding agents fourteen Jev judgment tools, turning evidence plus an enumerable question into auto, review, or escalate actions from calibrated probabilities.

Jev Agents & MCP#community#jevbest#catalog-import

Community

PyModel/jev-judge-mcpPython

jev-test-filter

Diff-aware test selector that sends the git diff to Jev with one question per test, then prints the native filter flags and hands them to vitest, jest, node, bun, Playwright, cargo, or go test.

Jev Agents & MCP#community#jevbest#catalog-import

Community

mizchi/jev-test-filterTypeScript

jevyoumean

Shell wrapper that intercepts unknown subcommands, reads the target CLI help output, and asks Jev which documented subcommand matches the intent behind the typo.

Jev Agents & MCP#community#jevbest#catalog-import

Community

syumai/jevyoumeanGo

jev

Claude Code plugin with jev find, jev ask, and a Read hook that keeps large-file context out of the agent by scoring relevance with Jev.

Jev Agents & MCP#community#jevbest#catalog-import

Community

BorisLeMeec/jevGo

jev-guard

Claude Code plugin that scores pending tool calls and finished answers with Jev noul probabilities, denying the call with the correction attached when a rule threshold is met.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

muratcakmak/jev-guardTypeScript

reflex-gate

Local Windows gateway that fronts agent loops with stop, extract, and scan endpoints plus a System One-shaped typed decision path. Runs a small coder model over llama.cpp at localhost with telemetry and self-tests.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

jagsan-cyber/reflex-gateGo

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai