jev-router
Auto-picks Haiku, Sonnet or Opus and effort per prompt, step and subagent.
Illustration. Actual results depend on the mod, terminal and environment.
Install
/plugin marketplace add dominicrico/jev-router
/plugin install jev-router@jev-router
/reload-pluginsRun these commands one at a time in a Claude Code session.
claude plugin marketplace add dominicrico/jev-router
claude plugin install jev-router@jev-routerThen run /reload-plugins in your session, or start a new session.
Please install the Claude Code mod "jev-router" (https://github.com/dominicrico/jev-router). Before installing, read the Security and Requirements sections of its README and tell me what it reads and writes, which commands it runs, whether it uses the network, and which Claude Code version, OS and terminal it needs. If everything looks fine, run claude plugin marketplace add dominicrico/jev-router and claude plugin install jev-router@jev-router, then remind me to run /reload-plugins in my session.Send this to Claude Code so it reads the author’s instructions before installing.
git clone https://github.com/dominicrico/jev-router
cd jev-router
claude --plugin-dir .Clone the repository and try it in a local session with --plugin-dir.
What it accesses
Based on the author’s documentation, checked when listed; use claude plugin validate to check for yourself.
Reads
YesReads prompts, optional recent messages, subagent task text and API key env vars.
Writes files
YesWrites only the Claude Code plugin store (including a plain-text /jev key).
Runs commands
UnspecifiedNetwork
YesSends routing requests to api.typesafe.ai or your configured Laya/Kev endpoint.
Introduction
jev-router is a Claude Code mod that asks TypeSafe Jev (or self-hosted Laya / Kev) which Claude model and reasoning effort fit each prompt, each step and each subagent, runs the turn on that pick, and shows it in a live band above the prompt. The default lean preset caps effort at medium; !full lifts it for one prompt and !opus / !sonnet / !haiku pin a model. A cache guard limits switches; after three failures it pauses a minute and keeps your model. The author's benchmarks claim large savings on small tasks but not in long warm sessions. Needs a TypeSafe API key. Prompts, optional recent messages and subagent task text go to api.typesafe.ai after redaction. No minimum Claude Code version is stated. Based on README and SECURITY; not run-tested by us.
Who it is for
- Opus-by-default users who want to save on small tasks
- People who want to see which model runs each turn
Requirements
- Needs a TypeSafe API key; optional with self-hosted Laya/Kev.