jev-cli MCP server
Commit Nov 16, 2026Allows AI agents to perform semantic classification and scoring on text by returning calibrated probabilities instead of conversational text.
Also known as shaharia-lab/jev-cli
- Stars
- 46
- Tools
- โ
- Downloads, 30 days
- โ
- Latest version
- โ
- Last commit
- Nov 16, 2026
- License
- Apache-2.0
What is jev-cli MCP?
jev-cli allows AI agents to perform semantic classification and scoring on text by returning calibrated probabilities instead of conversational text.
Summary generated from the project's README
About jev-cli MCP server
Automating content judgment through semantic classification is what jev-cli enables for developers and AI agents. Instead of parsing conversational text from large language models, this tool provides calibrated probabilities for specific questions. It acts as a semantic if statement, allowing users to branch logic in shell scripts, CI jobs, or automated pipelines based on numerical confidence rather than unpredictable prose.
Users can interact with the tool via a command-line interface or by running it as an MCP server using the jev mcp serve command. It requires a TypeSafe API key to function, which can be provided via an environment variable or a local authentication file. The tool offers three distinct question types: noul for yes/no properties, choice for selecting one of up to 255 options, and score for evaluating content against a custom scale. It is designed for high-efficiency tasks like triaging support tickets, moderating user content, or gating CI jobs based on commit descriptions.
What you can do with it
- 1Automated support ticket triageAn AI agent can determine which department should handle a specific ticket by asking a choice question.
- 2CI/CD pipeline gatingDevelopers can use the tool to decide if a commit requires a changelog entry based on its description.
- 3Content moderation and safetySystems can automatically flag user content that breaks specific rules while sending borderline cases to humans.
- 4Large scale feedback scoringA developer can process thousands of product reviews to determine sentiment levels in a single batch run.
Written from the project's README. Every use case is checked against a quote from it.
Configuration
TYPESAFE_API_KEY
Automated checks
Ran Nov 14, 2026Full report โNot verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work ยท What to fix for more points
- Source repository linked and reachableshaharia-lab/jev-cli is public and reachablegithub.com/shaharia-lab/jev-cli โPass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html โPass
- Commit or release in the last yearLast commit or release 2026-11-13github.com/shaharia-lab/jev-cli โPass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubAsk any text a typed question. Get a probability back, not a paragraph.
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 ยท What can I use it for? ยท Documentation ยท Command reference ยท FAQ
โญ Useful idea? Star the repo.
It takes two seconds, and it is how the next person finds jev.
๐ฌ See it in action
Ask a question, get a probability, branch on the exit code, then ask several questions at once.

[!IMPORTANT] Unofficial project.
jevis community-built. It is not affiliated with, endorsed by, or sponsored by TypeSafe AI. "TypeSafe" and "Jev" belong to their owner. You need a TypeSafe API key to use it.
๐ค What problem does this solve?
You have text: a support ticket, a commit message, a user's answer, a product review, a model's output. You need a decision about it: is it angry, which team owns it, how risky is it, should this pipeline continue?
Today you either write brittle keyword rules, or you ask a large language model and then parse a sentence it might phrase differently tomorrow.
Jev is a third option. You define the question and the possible answers. The model returns a number for each one. Your code decides what to do.
jev noul "Is this customer angry?" --state "You charged me twice. Fix it now."
answer noul yes 0.91 โโโโโโโโโโโโโโโโโโโโ
model jev-1.13.0 (requested jev-latest) ยท 42 input tokens ยท est. cost $0.000002 ยท 1 ms ยท request req_demo
One question, one number, about a hundredth of a cent. No prompt engineering, no JSON parsing, no "As an AI language model".
| LLM prompt | jev |
|
|---|---|---|
| You get back | A sentence to parse | A number in a shape you defined |
| Same input tomorrow | May be worded differently | Same contract, always |
| Cost of a short ticket | Cents | A fraction of a cent |
| Use in a shell script | Parse and hope | Exit code 0 or 10 |
๐ก What can I use it for?
Anywhere a program needs a judgment about content. A few that people reach for first:
| Use case | The question you ask |
|---|---|
| Triage support tickets | Which team owns this, how urgent is it, how upset is the customer |
| Route incoming messages | Sales, support, or spam |
| Moderate user content | Does this break the rule, with the borderline cases sent to a person |
| Gate a CI job | Does this commit describe a user-facing change, so it needs a changelog entry |
| Guard an AI pipeline | Is the user's input on topic, and did the model's answer actually answer it |
| Score feedback at scale | How positive is each of 50,000 reviews, as one batch run |
| Label a dataset | Turn a folder of documents into labelled rows, resumable if it breaks |
| Detect intent in a form | Is this person asking to cancel, and is a refund being requested |
| Check a document | Does this contract mention auto-renewal, in plain language |
Each of these is one command:
# Is this message angry? Exit 0 when yes is 0.7 or more, exit 10 when it is not
jev noul "Is this message angry?" --state-file message.txt --fail-under 0.7
# Which team should handle this ticket?
jev choice "Which team should handle this?" --state-file ticket.txt \
--option billing --option technical --option other
# How frustrated is this customer, on a scale you describe?
jev score "How frustrated is the customer?" --state-file ticket.txt \
--level Calm --level Frustrated --level "Very angry"
More ideas, and how to phrase a question so it works, in writing good questions.
โก Quick start
Three steps, about a minute, a fraction of a cent.
| Method | Command |
|---|---|
| Install script (Linux, macOS) | curl -fsSL https://raw.githubusercontent.com/shaharia-lab/jev-cli/main/install.sh | sh |
| Install script (Windows) | irm https://raw.githubusercontent.com/shaharia-lab/jev-cli/main/install.ps1 | iex |
| Homebrew (macOS, Linux) | brew install shaharia-lab/tap/jev |
| Cargo, prebuilt | cargo binstall jev-cli |
| Cargo, from source | cargo install jev-cli --locked |
The install scripts never use sudo, always check the download's SHA-256, and check its
minisign signature when minisign is installed. Homebrew
also sets up the man pages and shell completions. Options, verification and uninstalling are in
the installation guide.
Create one in the TypeSafe console, then either export it, which is what CI jobs and agents should do:
export TYPESAFE_API_KEY=...
or store it once, in a private file only you can read:
jev auth login
jev auth status
The key is never accepted as a flag value, so it cannot land in your shell history, and only its last four characters are ever shown.
jev noul "Is this message angry?" --state "You charged me twice. Fix it now."
Piped or redirected, the same command prints JSON instead of a table, and --field noul prints
0.91 and nothing else.
[!TIP] Ask several questions about the same content in one call. Extra questions cost tokens, not round trips, and the content is only paid for once.
โ The three kinds of question
| Type | Command | You get back | Ask it when |
|---|---|---|---|
| noul | jev noul |
The probability of yes, 0 to 1 | A property either holds or it does not |
| choice | jev choice |
The winning option, its confidence, and a probability for each | Exactly one of up to 255 options must win |
| score | jev score |
A position on your scale, and a confidence | A described scale of 2 to 10 levels |
A probability near 0.5 means the model cannot tell, not "medium". --abstain-band 0.4,0.6 turns
that into its own exit code, so those cases can go to a person. Jev reads literally: it cannot
count, do arithmetic or compare dates, so keep that in your code.
Details and worked examples: the three question types.
๐งฉ Many questions at once
One content, several questions, one call. Write triage.yaml:
model: jev-latest
questions:
is_urgent:
type: noul
instructions: Does this ticket convey urgency?
criteria:
"true": Explicitly time-sensitive
"false": No urgency expressed
department:
type: choice
instructions: Which team should handle this?
criteria:
billing: Payments, invoicing, refunds
technical: Bugs, outages, integrations
other: ~
frustration:
type: score
instructions: How frustrated is the customer?
criteria:
- Calm
- Frustrated
- Very angry
Check it for free, then run it:
jev validate -f triage.yaml --state-file ticket.txt
jev eval -f triage.yaml --state-file ticket.txt
jev eval -f triage.yaml --state-file ticket.txt --field answers.department.choice
jev validate sends nothing and needs no key. jev eval --dry-run goes further and prints the
exact request and what it should cost.
To run the same questions over thousands of rows, with concurrency, back-off and a --resume
that survives a crash, use jev batch run: see many rows at once.
๐ง Built for scripts, CI and agents
stdout is data. stderr is everything else. Exit codes are a stable contract, so a script never has to parse output:
| Code | Meaning |
|---|---|
| 0 | Success, and any gate condition holds |
| 2 | Usage or validation error. Nothing was sent, nothing was billed |
| 3 | No API key, or the API refused it |
| 10 | Evaluated, and the gate condition is false. Never an error |
| 11 | Evaluated, and the answer is inside the abstain band |
if jev noul "Is this ticket about billing?" --state-file ticket.txt --fail-under 0.7 --quiet; then
echo "billing"
fi
The full table, gating flags, JSON shapes, the error object and a GitHub Actions example are in scripting and CI and exit codes and JSON contract.
๐ค For AI agents
jev treats agents as first-class users, not an afterthought.
- Self-describing.
jev specprints every command, flag, default, exit code and example as one JSON document, and every--helpsays when to use this command rather than its siblings. - Schemas, not guesswork.
jev schema request|questions|batch-record|output|errorprints JSON Schemas for everythingjevreads and writes. - Free dry runs.
jev validateand--dry-runcheck a request offline, so an agent can fix its own mistake before spending anything. - MCP server.
jev mcp serveexposes the same functionality as MCP tools, with a per-call spend cap and file access that is deny-by-default.
claude mcp add jev -- jev mcp serve
There is also an agent skill that teaches the whole workflow:
claude plugin marketplace add shaharia-lab/jev-cli
claude plugin install jev@jev-cli
Setup for Claude Desktop, Cursor, VS Code and any other MCP client, plus the spend caps, is in AI agents and MCP.
๐ Privacy and security
- No telemetry, ever.
jevcontacts exactly two hosts: the TypeSafe API when you evaluate something, and GitHub Releases when it checks for an update, which you can turn off. - Your content stays yours. It goes to the API and nowhere else, and is never written to a log unless you explicitly ask for that.
- The key is never exposed. Not as a flag, not in logs, errors, dry-run output or MCP results. Only its last four characters are ever shown.
- HTTPS only, certificate verification cannot be disabled, and redirects are never followed.
- Signed updates. Every download is checked against a signature and a checksum before
anything is replaced, and
jevnever downgrades on its own.
SECURITY.md has the full model, how to verify a release yourself, and how to report a vulnerability privately. Please do not open a public issue for one.
๐ Documentation
User guide
- Installation
- Quick start
- The three question types
- Writing good questions
- Scripting and CI
- Exit codes and JSON contract
- Many rows at once
- Configuration and profiles
- AI agents and MCP
- Keeping jev up to date
- Troubleshooting
- FAQ
Reference and development
- Command reference, generated from the binary
- Development setup
- Architecture
- Testing
- Releasing
- How the API actually behaves
- Threat model
jev --help # every command
jev spec # the whole contract, as JSON
jev schema request # what a request file may contain
๐ FAQ
An LLM writes a sentence you then have to parse, and it may phrase things differently tomorrow. Jev returns a number in a shape you specified, so your code can branch on it. It is also much faster and much cheaper, because it generates nothing. It is not a replacement for an LLM: it is the piece you reach for when the job is a decision.
Only input tokens are billed, so a short ticket costs a small fraction of a cent. jev always
labels a cost as an estimate, and --dry-run prices a run before you send it.
Yes, for anything that evaluates content. Everything offline is free and needs no key:
jev validate, jev schema, jev spec, jev completion and any command with --dry-run.
Not bit for bit. Compare against thresholds, never for equality, and pin a versioned model such
as jev-1.13.0 for anything that must stay stable. The alias jev-latest moves to new versions
without notice.
Yes. jev batch run streams JSONL or CSV, so memory does not grow with the file, backs off when
the API rate limits, writes one JSON record per row, and picks up where it stopped with
--resume.
In TYPESAFE_API_KEY, or in a private credentials file with mode 0600. There is no OS keychain
and no browser flow, by design.
More answers: the full FAQ and troubleshooting.
โญ Spread the word
jev is a young project around a young model. The fastest way to keep it alive is to make it
easier for the next person to find.
โญ Star jev ยท ๐ Report something broken ยท ๐ฌ Share what you built
๐ค Contributing
Pull requests are welcome. CONTRIBUTING.md explains the workflow and the quality gates, and CLAUDE.md is the architecture guide that contributors and AI agents both work from.
git clone https://github.com/shaharia-lab/jev-cli
cd jev-cli
make check # everything CI runs
Security problems go privately through SECURITY.md, never a public issue.
Licence
Licensed under either of Apache License, Version 2.0 or MIT license at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this project by you, as defined in the Apache-2.0 licence, shall be dual licensed as above, without any additional terms or conditions.
Built by Shaharia Lab ยท powered by TypeSafe AI
โญ Star jev if it saved you a prompt.
jev-cli MCP questions
Is it compatible with AI agents like Claude?
Yes, it includes an MCP server mode that can be added to Claude Desktop, Cursor, or VS Code to expose its functionality as tools.
Do I need an API key?
Yes, you need a TypeSafe API key to use the tool.