# Jev MCP

Vendor: Jev  
Category: other  
Auth: API_KEY  
Composio slug: `jev`

Jev is TypeSafe AI's System One API for discovering available models and evaluating structured state with Noul, Choice, and Score questions.

## Verdict

Jev runs structured evaluations on shared state — text, data, or conversation history — using named questions you've defined in your Jev account. When you @mention Jev in a Project, it scores or classifies that state against your custom rubrics (Noul, Choice, or Score questions) in parallel. This is useful for teams that need consistent judgment calls on content quality, compliance checks, or decision criteria. Each evaluation consumes inference credits in your Jev account, so you'll want to be deliberate about when you invoke it. Best for product teams running repeatable assessments, not ad-hoc brainstorming.

## Tools

- **Evaluate State** — Evaluate one shared state against one or more named Noul, Choice, or Score questions in parallel; each call performs usage-priced inference.
- **List Models** — List the Jev model names and aliases available to the connected account for use in evaluations.

## Setup

1. Create or open a Switchy Project.
2. Open Integrations and connect Jev.
3. Authorise through api_key.
4. Teammates can now @-mention Jev tools in chat.

---
Last verified: 2026-10-17T02:00:10.360Z  
Canonical URL: https://switchy.build/directory/mcps/jev