Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

jev-rerank

Agent skill and Python library for Claude Code and Codex that reranks candidate lists with one Jev yes/no relevance question each.

Projectbeta
View on GitHub ↗

Overview

This jev-rerank reorders a query's candidate list using Jev's Noul primitive, following the approach in TypeSafe's own rerank cookbook. You bring a shortlist from BM25, grep, embeddings or anything else; for each candidate Jev answers whether it could be the right one for the query, and results are sorted by that probability. Every call is independent, so no candidate's score depends on another.

The repository is itself an agent skill: SKILL.md sits at the root, so cloning it into Claude Code's or Codex's skills directory makes it available, and the agent then uses the library instead of improvising API calls when asked to rerank search results or support articles. It also works as a standalone Python library and CLI. A one-time configure step stores a key for OpenRouter, TypeSafe direct or a custom endpoint. It is an independent community client.

What the decision model decides

TypeQuestionAnswers
Yes / noCould this candidate be the right one for this query?yes · no

How it works

  1. Clone into the agent's skills directory and install
  2. Run jev-rerank-configure once
  3. Ask the agent to rerank candidates, or call rerank() directly
  4. Candidates come back sorted by probability

Strengths

  • Works as an agent skill with no build step
  • Independent per-candidate scoring
  • Supports OpenRouter, TypeSafe or a custom endpoint

Caveats

  • One call per candidate can add cost for long lists
  • Community client, not official

Getting started

git clone https://github.com/openwhat007/jev-rerank ~/.claude/skills/jev-rerank && cd ~/.claude/skills/jev-rerank && uv pip install -e ".[dev]" && jev-rerank-configure

At a glance

Type
Project
Status
beta
Language
Python
Runtime
Agent skill, library and CLI
Has code
yes

Use cases and integrations

How curated lists describe it

Frequently asked questions

What does Jev decide?

Whether each candidate could be the right answer for the query.

How do I use it in Claude Code?

Clone it into ~/.claude/skills/jev-rerank, install, configure, then ask Claude to rerank.

Which endpoints work?

OpenRouter, TypeSafe direct or a custom endpoint.

Related projects

  • hippo-memory

    Local long-term memory for coding agents with an optional Jev reranker that orders recalled memories.

    ProjectproductionTypeScript
  • ReadyCode Reader

    Local-first MCP server that returns only Jev-checked passages from big PDFs, Word files and spreadsheets, with exact row calculations.

    MCP serverbetaTypeScript
  • siftr

    CLI and MCP server giving coding agents Jev-powered semantic search, focused reads and list picking in about a second for about a cent.

    BenchmarkbetaPython
  • jevex

    Single MCP tool that indexes a repo and has Jev rank a shortlist so a coding agent reads the right files and lines instead of grepping.

    MCP serverbetaPython
  • jev-filter

    npm CLI that screens large agent tool outputs with Jev, returning compact typed decisions and IDs while originals stay recoverable.

    CLIbetaTypeScript
  • apple-rag-mcp

    MCP server over Apple developer docs and WWDC transcripts that uses hybrid search, then Jev relevance scores to rank results.

    MCP serverproductionTypeScript