jevbooks

← All projects

jev-palette

amycardoso/jev-palette

Semantic command palette ranked by intent with TypeSafe Jev.

What it is

A demo command palette (⌘K) for a fictional mail app. It ranks 77 commands by meaning from Portuguese or English intent, side by side with a classic fuzzy matcher that fails on the same queries.

How it uses Jev

One Choice question per keystroke over the whole 77-command catalog as options. Jev returns the full probability distribution, which is used directly as the ranking. The selected command is then executed by deterministic code.

Primitives:choice

Technique worth stealing

Use the full probability distribution from a single Choice call as the ranking, avoiding embeddings and vector indexes.

Try it

npm install; cp .env.example .env; add TYPESAFE_API_KEY; npm run dev. Without a key, npm run dev:mock.

View on GitHub

judged by Jevjev-1.13.0

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-centricyes0.97
  • Shows a System One patternyes0.87
  • Handles uncertaintyno0.05
  • Measuredyes0.67
  • Runnableunclear0.50
  • Worth recommendingyes0.75
  • Model replicano0.03
  • Problem scopescore on a 0–2 scale0.15
  • About Jevyes0.99

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.