clarity-judge
Multi-axis writing quality checker powered by TypeSafe AI's Jev model.
What it does
Multi-axis writing quality checker powered by TypeSafe AI's Jev model. Separate named checks, each with its own verdict and confidence. Sandbox first, invoice later: poke at Choice, Score, and Noul before you wire billing. Tags in the wild: playgrounds.
Skim the repo for setup, then open the demo when you want proof before you merge anything.
How Jev is used
Runs named checks (each a System One question) against pasted writing, surfacing per-check verdicts and supporting sentences instead of one opaque LLM grade.
Setup: Clone TypeSafeAI/clarity-judge for local runs with your API key. Public judge.jev.works demo is deterministic simulation only.
Why multi-axis beats one score
Editors can accept good tone while rejecting passive voice. Separate Jev questions keep each axis thresholdable and explainable, similar to the jev.works playground UX.
FAQ
- Does the public demo call Jev?
No. README states judge.jev.works has no server API key and shows simulated results labeled as demo mode.
- Is this an application or playground?
It is filed under playgrounds/apps paths because it is an interactive UI for experimenting with check definitions, even though it targets writing quality use cases.
Tags
Related products
Hand-picked neighbors with rich profiles or overlapping tags.