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

jev-router (dominicrico)

Claude Code plugin that routes each prompt, step and subagent to the cheapest capable Claude model with effort caps and a live band.

BenchmarkbetaMIT
View on GitHub ↗Live demo ↗

Overview

jev-router (dominicrico) is a free Claude Code plugin that sends each prompt, step and subagent to the cheapest Claude model that can do the job (Haiku, Sonnet, Opus or Fable), caps runaway reasoning effort, and shows every pick in a live band above the prompt. A lean preset is on by default, so setup is just installing the plugin and providing a TypeSafe key.

Features include re-routing at every step, routing subagents, an effort cap, an optional Sonnet ceiling, a prompt-cache guard, per-prompt control and privacy controls. The README presents benchmarks from real Claude Code runs, with hard tasks graded by hidden tests, measured against always using Opus: about 97% lower cost on small tasks, 65% on everyday work, 61% on hard tasks with 39 of 40 completed, and 56% on agent runs with subagents with 18 of 18 completed, at about 253 ms per pick. It also notes where routing does not help, saying hard tasks can cost more.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich Claude model should handle this prompt, step or subagent?haiku · sonnet · opus · fable
ScoreWhat reasoning effort does this need?(inferred)—

How it works

  1. Install the plugin and set the TypeSafe key
  2. On each prompt, ask Jev for a model and effort
  3. Check the cache guard and apply effort caps
  4. Re-ask each step and route subagents
  5. Show the pick in the live band

Strengths

  • Large reported savings versus always-Opus
  • Benchmarks graded by hidden tests
  • Live visibility of every pick

Caveats

  • Some hard tasks can cost more, per the README
  • Requires a TypeSafe key
  • Benchmarks are the author's own runs

Getting started

/plugin marketplace add dominicrico/jev-router, then /plugin install jev-router@jev-router and /jev key <your TypeSafe key>

At a glance

Type
Benchmark
Status
beta
Language
TypeScript
Runtime
Claude Code plugin
Has code
yes
License
MIT

Use cases and integrations

How curated lists describe it

  • Claude Code plugin: automatic Claude model routing.

    wh000wh000/awesome-jev-live

  • Claude Code plugin: automatic Claude model routing. Picks Haiku, Sonnet or Opus and the reasoning effort per prompt, step and subagent; effort cap, prompt-cache guard, status band, benchmarks.

    hellogumbo/awesome-jev

Frequently asked questions

How much does it save?

The README reports about 97% on small tasks and 56-65% on typical and agent work versus always using Opus.

Does it route subagents?

Yes, as well as each prompt and step.

Does quality suffer?

In the README's hard-task benchmark, 39 of 40 tasks were completed.

Related projects

  • jev-router★ 593

    Per-turn model routing for Claude Code and Codex: wraps the real CLI and lets Jev send simple turns to fast models and hard ones to strong.

    ProxybetaJavaScript
  • Claude Model Router

    Claude Code mod that sets model and effort per turn and subagent from a prompt classifier: Jev by default, or Clef, OpenAI or Ollama.

    Proxybeta
  • coding-router-jev

    TypeScript/Bun launcher that asks Jev which model and reasoning effort each Codex, Claude Code or Pi turn should use.

    ProjectbetaTypeScript
  • decision-router

    CLI, Claude Code plugin and Pi virtual model that pick the right model per prompt from one Jev call with a complexity floor.

    PluginbetaTypeScript
  • effortless

    Claude Code mod that picks reasoning effort per prompt with a Haiku judge, optionally Jev, and shows cache and context state.

    ProjectbetaTypeScript
  • frost

    Go CLI model router that uses one Jev analysis to recommend a model, harness and effort level, then selects deterministically.

    CLIbetaGo