What it is
A local-first CLI that picks an AI coding agent, model, and reasoning effort for a coding phase, then launches it in a Herdr pane. Built for developers managing multiple agent subscriptions and shared accounts.
How it uses Jev
Fixed rules filter eligible models by quota, policy, and the 40% reserve on shared accounts. TypeSafe then ranks the remaining candidates and selects an effort for the task phase; the winning agent, model, and effort are launched in a Herdr pane to execute the task.
Primitives:choice
Technique worth stealing
Deterministic eligibility filtering before semantic ranking, so quota and policy rules constrain the candidate set passed to TypeSafe.
Try it
npm install && npm run build, then router run "<task>" --dry-run
Evidence
Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.
- Jev-centricunclear0.53
- Shows a System One patternunclear0.44
- Handles uncertaintyno0.04
- Measuredno0.04
- Runnableyes0.97
- Worth recommendingno0.29
- Model replicano0.04
- Problem scopescore on a 0–2 scale1.20
- About Jevyes0.96
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.