An MCP server built around measuring questions rather than asking them: jev_ask prototypes wording, jev_eval sweeps thresholds against labeled data, jev_map applies the winner over a workload.
User repository on GitHub
An eval-first MCP server for TypeSafe's Jev, a System One model that returns typed judgments (noul, choice, score) with probabilities instead of generated text.
Repo stats from the GitHub API, cached Oct 2026.
This is one of several Jev listings for the same project. The canonical entry is Jev MCP — jkudish.
An eval-first MCP server for TypeSafe's Jev, a System One model that returns typed judgments (noul, choice, score) with probabilities instead of generated text. Most Jev integrations expose "ask the model a question". The hard part in practice is not asking — it is knowing whether a question works and where to put the threshold. This server is built around that:
Quoted from the project's README. Read the full README.
Watch out
You need Node 20 or newer and a TypeSafe API key, and the quickstart runs the server through npx -y, which fetches the published package rather than anything you have read; build from a clone and point your client at dist/index.js if you would rather pin the code. Question ids are yours alone and are not transmitted, so all of the meaning has to live in the instructions string or the criteria. The evaluation tools are only as good as the labeled data you supply, and thresholds are tied to a specific model version, which is why jev_eval prints the version it resolved. The sample run in the README uses 16 labeled items and two variants, which demonstrates the output format rather than establishing accuracy on your data. Copy reviewed against the repository README on 2026-10-04, not from running it.
nassim-arifette/jevgrepTypeScript
opencx-labs/zevalsTypeScript
rashedInt32/jev-mcpTypeScript
BeLazy167/typesafe-modTypeScriptPosts from builders shipping with Jev right now.
