Is this just a thin curl wrapper I could write in an afternoon?

Mostly, but the ergonomics are the point. Questions live in text or JSON files you can put under version control, --context-json passes structured state, several questions share one call against that context, and named questions print as name=answer. That combination is what makes it usable inside a pipeline.

How do I tell a confident answer from a coin flip?

Two distinct numbers. --distribution prints the probability of every option, while --stats prints confidence, which the README defines as how sure the model is that it can answer the question with those options supplied. It recommends checking confidence, since a low value means being cautious about acting.

What happens to my script when the model is unsure?

Set --min-confidence and that question prints an empty answer, a note goes to stderr, and the run exits 2 after every line has printed. Partial results still land on stdout, so you have to check the exit code before reading output, which is a real footgun in shell conditionals.

Will the same question get the same answer with different options?

No, and the README walks through the case. Given Atlanta and Chicago for what the capital of Georgia is, it answers Atlanta at confidence 1.000. Add Tbilisi as an option and confidence falls to 0.520 while Atlanta drops to 0.240. The option set is part of the question.

How do I install and configure it?

Run brew install vsekhar/tap/decide, then decide --set-config --model typesafe:jev-latest --api-key - and paste the key at the prompt. You can also set DECIDE_MODEL and DECIDE_MODEL_API_KEY in the environment, or write them to $HOME/.config/decide/config.

How do I pass structured or named context?

Use --context-json to parse a value as JSON, for example an order object with total and days_since_delivery, or --context refund_policy=@refund_policy.txt to name a file and refer to it by name in the question and option text.

What can a JSON question file express?

Names, richer descriptions and structured instructions: options can carry a summary, examples, signals and not_for, levels can carry summaries, and a question can set min-confidence and a fallback. Text question files mimic the command line, and both can be mixed with inline questions.

CommunityJev Integrations & Adapters

Decide — vsekhar

A shell-first CLI for yes or no, multiple-choice, and leveled decisions through TypeSafe Jev, composable with files and JSON context in scripts.

See highlights

Open github.com

Stars: 2
Language: Swift
License: APACHE-2.0
Published by: Community
Author: vsekhar
Last verified: Oct 2026

vsekhar/decide

User repository on GitHub

View on GitHub

Make decisions from the command line, in scripts, and in agent skills. No code required.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Make decisions from the command line. No code required. Use decision models like Typesafe's Jev to make smart decisions in your scripts, pipelines, and agent skills. Decision models let you ask and answer yes/no, multiple choice, and multiple level questions. Decision models can return an answer, the model's confidence in its answer, the probability of the answer being true, and even a full distribution across possible answers.

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

Tagscommunity cli shell agents

Works forJev for Classification Jev for Routing Jev for Scoring

Used inJev for Software & DevTools

Languages: Go
Platforms: CLI
Capabilities: Open sourceRequires API key

Highlights

brew install vsekhar/tap/decide; configure typesafe:jev-latest and an API key interactively or via DECIDE_MODEL env vars.
Supports --context @file.txt, named contexts, and --context-json for structured state in one invocation.
Multiple questions against the same context can ship in one call per the README examples.

Quickstart

bash

decide --context @ticket.txt \
  "Which team handles this ticket?" \
  --option shipping --option billing --option returns

Watch out

Needs a TypeSafe key for hosted Jev. Agent-skill examples are community patterns, not bundled enforcement.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

Hacker News: Show HN for the decide CLI: brew install, file context, and multi-question shell scripts.

vsekhar2 pointsShow HN: Decide – Jev decisions in the shell, scripts, and agent skillsRead the thread on Hacker News

Other listings of this project

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

decide

pydecide is one Python client for typed System One decisions — Choice, Score and Noul — over hosted Jev, OpenRouter's Decisions endpoint, local laya on PyTorch or MLX, a sentence-transformers CrossEncoder, or a JSON-prompted LLM fallback.

Jev Practices & Patterns#community#awesomejev#catalog-import

Community

gopaljigaur/decidePython

More Jev resources like this

onesie

A Unix-pipeable CLI that asks Jev (or Berget System One models) one question over stdin, with shell-safety exit codes and a calibrate subcommand.

Jev Integrations & Adapters#community#cli#shell

Community

frodi-karlsson/onesieGo

Jev-Mail

CLI-installed Gmail classifier that runs 24/7 on user-owned Google Apps Script: the worker asks Jev for three typed judgments per thread (requires_action, is_important, bucket), applies labels, stars, and optional archiving, and routes ambiguous or low-confidence mail to a Review label.

Jev Integrations & Adapters#community#gmail#email

Community

vynnlee/jev-mailTypeScript

AgentScope

33k

Alibaba's multi-agent platform, which implements native TypeSafe Jev classification through JevClassifierModel for binary, choice and score routing across agent pipelines. Python 3.11+, Apache-2.0, with a FastAPI agent service and web UI.

Jev Integrations & Adapters#community#multi-agent#framework

Community

agentscope-ai/agentscopePython

Paca

1.9k

A Go-based team workspace with optional Jev AI for agent auto-routing, task auto-fill, auto-assign, and automation condition nodes.

Jev Tools & CLIs#community#go#agents

Community

Paca-AI/pacaGo

OpenRouter AI SDK provider

689

The Vercel AI SDK provider for OpenRouter, including openrouter.decisionModel for typesafe/jev-1.13 evaluations with providerMetadata answers.

Jev Integrations & Adapters#community#typescript#openrouter

Community

OpenRouterTeam/ai-sdk-providerTypeScript

JIT-JEV Context OS

Epistemic runtime for autonomous coding agents that uses Jev as a sub-100ms System 1 gate to shadow-evaluate capsule facts, propose evictions before bloat and prefetch memory between turns. Its fixture run reports 31.3% fewer turns and 52.6% fewer blind discovery operations, with a fail-open circuit breaker.

Jev Integrations & Adapters#community#agents#context-engineering

Community

wojciechwiesner/jit-contextPython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai