jev-mcp
Python FastMCP server wrapping Jev's noul, choice and score judgments, with configs for many coding agents and an OpenCode Zen route.
Overview
This jev-mcp is a Python MCP server built on FastMCP that wraps TypeSafe's Jev System One models so agents get typed judgments instead of prompt-and-parse. It exposes jev_evaluate for a map of typed questions in one parallel call, jev_noul for yes/no probabilities, jev_choice for picking one option from a rubric map, jev_score for 2-10 ordered levels, and jev_models to list available models.
Every tool accepts state as a string, object or array, an optional model (default jev-latest) and an endpoint of direct or zen. The zen route uses an OpenCode Zen key for jev-1.13 or jev-1.13-free. The README gives ready configurations for OpenCode, Claude Code, Codex CLI, Gemini CLI, Cursor and Windsurf.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Yes/no judgment on the state (0 = no, 1 = yes) | yes · no |
| Choice | Pick one option from the criteria map | — |
| Score | Rate the state along 2-10 ordered levels | — |
How it works
- Clone the repo and add TYPESAFE_API_KEY to .env
- Run with uv run --with fastmcp --with httpx server.py
- Register in an MCP client using the provided config
- The agent calls a tool and branches on the typed answer
Strengths
- Ready configs for six coding agents
- Optional OpenCode Zen endpoint including a free model tier
- Single call evaluates several questions in parallel
Caveats
- AGPL-3.0 licence
- Requires a TypeSafe key (or OpenCode Zen key for zen)
- Thin wrapper over the API
Getting started
git clone https://github.com/rajasekharponakala/jev-mcp.git && cd jev-mcp && cp .env.example .env && uv run --with fastmcp --with httpx server.pyAt a glance
- Type
- MCP server
- Status
- beta
- Language
- Python
- Runtime
- MCP server (FastMCP)
- Has code
- yes
- License
- AGPL-3.0
Use cases and integrations
How curated lists describe it
MCP server wrapping TypeSafe's Jev System One models — typed noul/choice/score judgments for AI agents
logicrw/awesome-jev-projectsMCP server wrapping TypeSafe.
wh000wh000/awesome-jev-liveMCP server wrapping TypeSafe's Jev System One models — typed noul/choice/score judgments for AI agents.
hellogumbo/awesome-jev
Frequently asked questions
What tools does it provide?
jev_evaluate, jev_noul, jev_choice, jev_score and jev_models.
What is the zen endpoint?
An alternative route using an OpenCode Zen key with jev-1.13 or jev-1.13-free.
Which agents are documented?
OpenCode, Claude Code, Codex CLI, Gemini CLI, Cursor and Windsurf.
Related projects
CLI, MCP server and Claude Code plugin for closed-form yes/no, choice, score and rank decisions via Jev, OpenAI Decisions or Claude.
MCP serverbetaTypeScriptCLI and MCP tool that finds hard-coded judgment calls in your code, scores how well Jev fits, and replaces strong fits with a fallback.
ProjectbetaTypeScriptAgent skill and CLI that converts plain requirements, LLM prompts or prompt-calling code into a typed Jev request and runnable script.
Agent skillbetaPythonPyPI MCP server for Jev with choice, score, noul, compare, verify and batch classify tools plus a one-command client installer.
MCP serverbetaPythonMCP server exposing Jev as a single evaluate tool for Claude Code, Codex, OpenCode, Antigravity or Cursor.
MCP serverbetaRustMCP server, CLI and Claude Code skill with Choice, Score and Noul primitives and a confidence gatekeeper, hosted or offline.
MCP serverbetaTypeScript