cd /news/ai-tools/stop-parsing-llm-answers-classify-an… · home › topics › ai-tools › article
[ARTICLE · art-140600] src=dev.to ↗ pub= topic=ai-tools verified=true sentiment=↑ positive

Stop Parsing LLM Answers: Classify Any Text From Your Terminal With jev-cli

A developer has released jev-cli, an open-source command-line tool that wraps TypeSafe AI's Jev model to return calibrated probabilities instead of prose, letting scripts branch on typed answers via exit codes. The tool supports three query types — noul (yes/no probability), choice (multi-option selection), and score (rubric-based rating) — and can run as an MCP server so AI agents can offload small classification judgments to it. The project is community-built and not affiliated with TypeSafe AI.

by read4 min views1 publishedSep 27, 2026

Every developer who has plugged an LLM into a script knows this moment.

You ask a simple question: "Is this customer angry?" The model answers with a friendly paragraph. Now you write code to parse that paragraph. Tomorrow it words things differently, and your parser breaks.

You didn't want a paragraph. You wanted a yes or no, and ideally how sure it is.

That is why we built jev-cli, a free, open-source command-line tool that turns any text into a clear, typed answer your code can act on.

jev is a command-line tool for TypeSafe AI's Jev model, the model that never writes text Send it some content and a question you defined, get a calibrated probability, and branch on it in a shell script, a CI job or an AI agent. It is a semantic if statement for your terminal.

Quick start · What can I use it for? · Documentation · Command reference · FAQ

It takes two seconds, and it is how the next person finds jev.

Ask a question, get a probability, branch on the exit code, then ask several questions at once.

Important

Unofficial project. jev is community-built. It is not affiliated with, endorsed by, or sponsored by TypeSafe AI. "TypeSafe" and "Jev" belong to their owner. You…

jev-cli is an open-source CLI for Jev, a model from TypeSafe AI that never writes text. Instead of a reply, it gives you back a probability.

You give it two things:

You get back a number. Your script decides what to do with it.

jev noul "Is this customer angry?" --state "You charged me twice. Fix it now."
answer  noul  yes  0.96
model jev-1.13.0 · 280 input tokens · est. cost $0.000012

That's it. No prompt engineering, no JSON parsing, no "As an AI language model...".

Think of it as an if statement that understands meaning.

Type Ask it when You get back
noul Something is true or not Probability of "yes", 0 to 1
choice One option out of many must win The winner plus a probability for each option
score You want a position on a scale A level on your own 2 to 10 step rubric

For example, routing a ticket to the right team:

jev choice "Which team should handle this?" --state-file ticket.txt \
  --option billing --option technical --option other
Regular LLM prompt jev-cli
What you get A sentence to parse A number in a shape you defined
Same input tomorrow Might be worded differently Same contract, every time
Cost per short ticket Cents About a thousandth of a cent
In a shell script Parse and hope Branch on the exit code

Answers become exit codes, so a shell script never has to read output:

if jev noul "Is this ticket about billing?" --state-file ticket.txt --fail-under 0.7 --quiet; then
  echo "send to billing"
fi

Exit 0 means the condition holds. Exit 10 means it doesn't (and that is never treated as an error). Anything unsure can land in its own "ask a human" band with --abstain-band 0.4,0.6.

If you use Claude Code, Cursor or another AI agent, jev-cli can make it faster and cheaper. Instead of the big model spending tokens deciding "is this relevant?" a hundred times, it hands those small judgments to jev.

jev mcp serve runs it as an jev spec and jev schema describe every command in JSON, so agents don't guess--dry-run checks a request offline, for free, before anything is spent

claude mcp add jev -- jev mcp serve
curl -fsSL https://raw.githubusercontent.com/shaharia-lab/jev-cli/main/install.sh | sh

brew install shaharia-lab/tap/jev

Add your TypeSafe API key and ask your first question:

export TYPESAFE_API_KEY=...
jev noul "Is this message angry?" --state "You charged me twice. Fix it now."

Windows, Cargo and other options are in the installation guide.

jev-cli is free, open source and built in the open. If it saves you a parsing headache or a few dollars of tokens:

Disclosure: we maintain jev-cli. It is an unofficial, community project and is not affiliated with or endorsed by TypeSafe AI. "TypeSafe" and "Jev" belong to their owner. You need a TypeSafe API key to use it.

── more in #ai-tools 4 stories · sorted by recency
── more on @jev-cli 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/stop-parsing-llm-ans…] indexed:0 read:4min 2026-09-27 · —