Skip to content

jev-mcp-server

PyPI MCP server for Jev with choice, score, noul, compare, verify and batch classify tools plus a one-command client installer.

MCP serverbetaMIT

Overview

jev-mcp-server is a Python MCP server for TypeSafe's Jev that exposes the three official question types (choice, score, noul) plus compare, verify and batch classify. A one-command installer writes the client configuration for Claude Code, Pi, Cursor, OpenCode or Codex.

The README explains Jev as a hosted decision model that returns a typed answer with calibrated probabilities and a confidence, without prose or explanations, and states a single call costs roughly $0.00002 to $0.0001 and returns in about half a second. Choice supports 2-100 options and returns probabilities over all of them so near-ties are visible; score grades on a 2-8 level rubric and returns a fractional index; noul gives a 0-1 degree. A setup tool can verify a key live and store it with restricted permissions instead of placing it in client config.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich of these rollout plans is safest?—
ScoreGrade risk, severity or quality on an ordered 2-8 level rubric—
Yes / noIs this change breaking?yes · no
Choicecompare: which of A or B is better?A · B

How it works

  1. Get a TypeSafe API key
  2. Run uvx jev-mcp-server install for your client
  3. Restart the client and ask a decision question
  4. The agent calls a tool and sees the probability split

Strengths

  • Installer writes client config automatically
  • Near-ties are visible because all option probabilities are returned
  • No local model or second LLM required
  • Published on PyPI with CI

Caveats

  • Hosted dependency on TypeSafe
  • Cost and latency figures are the README's own estimates

Getting started

uvx jev-mcp-server install claude-code (or pi, cursor, opencode, codex)

At a glance

Type
MCP server
Status
beta
Language
Python
Runtime
MCP server (uvx)
Has code
yes
Has tests
yes
License
MIT

Use cases and integrations

How curated lists describe it

  • MCP server for Jev (TypeSafe System One): the three official question types — choice, score, noul — plus batch classify. Calibrated probabilities, ~0.5s, \<$0.001/call.

    logicrw/awesome-jev-projects
  • MCP server for Jev (TypeSafe System One): the three official question types…

    wh000wh000/awesome-jev-live
  • MCP server for Jev (TypeSafe System One): the three official question types — choice, score, noul — plus batch classify. Calibrated probabilities, ~0.5s, <$0.001/call.

    hellogumbo/awesome-jev

Frequently asked questions

Which tools does it expose?

choice, score, noul, compare, verify and batch classify.

How do I install it?

Run uvx jev-mcp-server install followed by your client name.

What does score return?

A fractional index on the rubric, for example 1.88 means between levels 1 and 2, leaning to 2.

Related projects

  • Gives Claude Code, Codex, Pi and OpenCode five typed Jev tools (classify, check, score, rank, ask) plus a skill on when to use them.

    MCP serverbetaTypeScript
  • MCP server giving any agent Jev's classify, score, yes/no check and allow/confirm/block gating of risky tool calls.

    MCP serverbetaJavaScript
  • Python FastMCP server wrapping Jev's noul, choice and score judgments, with configs for many coding agents and an OpenCode Zen route.

    MCP serverbetaPython
  • MCP server exposing Jev as a single evaluate tool for Claude Code, Codex, OpenCode, Antigravity or Cursor.

    MCP serverbetaRust
  • MCP server, CLI and Claude Code skill with Choice, Score and Noul primitives and a confidence gatekeeper, hosted or offline.

    MCP serverbetaTypeScript
  • Agent skill and CLI that converts plain requirements, LLM prompts or prompt-calling code into a typed Jev request and runnable script.

    Agent skillbetaPython