Jev in production › Model and agent routing

jev-router (dirien)

Local proxy asks Jev to pick a fast, balanced or frontier model tier for each human turn in Claude Code and Codex CLI sessions.

Open on GitHub ↗

The source publishes no measured number.
Use
Model and agent routing
Industry
Developer tools
Form
Open-source tool
Stage
Beta
Plugs into
Claude Code, Codex CLI
Listed
2026-10-16
Found via
github
Repository
dirien/jev-router
Stars
17
Forks
4
Last push
2026-10-29
Language
JavaScript
License
Apache-2.0
jev-router (dirien) screenshot
docs/live-view.png in the dirien/jev-router README, Apache-2.0; shown from GitHub.

The README opens with

A local model router for Claude Code and the Codex CLI that picks a model tier for each human turn with Jev.

jev-router is a pass-through proxy that runs on your own Mac or Linux machine. Claude Code talks to it in Anthropic Messages and Codex in OpenAI Responses. When you write a new message, the router asks Jev, TypeSafe AI's System One decision model, which tier the work needs, and sends the session to a fast, balanced or frontier model.

Many turns in a coding session don't need the most expensive model. Switching models in the middle of a task costs more than it saves, though: the prompt cache is per model, and a stronger model that inherits a weaker model's half-done transcript recovers only part of the quality gap. So the router decides...

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for model and agent routing