component-charades
Taboo-style browser game where Jev guesses which of 135 UI components you are describing and buzzes at 0.85 confidence.
Overview
Component Charades is a parlour game for design-system people. You are dealt a UI component card with forbidden words and must describe it until Jev guesses it. As you type, the app sends one request after each short pause and Jev re-ranks every component in the library, showing a live probability distribution. Below 0.5 it is listening, between 0.5 and 0.85 it is leaning, and at 0.85 it commits.
A second mode, Find a component, takes a job description and returns the component that does it, how sure Jev is, and separately whether the library has anything for it at all. The page layout mirrors a Jev request (STATE, CHOICE distribution, CONFIDENCE). The README cites 150 to 400 ms and about $0.0002 per request and links a write-up on what Jev is bad at and what was kept in ordinary code.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which component is being described (which_component)? | — |
| Yes / no | Does the library contain a component for this job? | yes · no |
How it works
- Player types a description (the state)
- After 350 ms of quiet, one Jev Choice request ranks all components
- Odds board redraws from the distribution
- At 0.85 confidence Jev buzzes; score is the number of words used
Strengths
- Shows live calibrated distributions in an intuitive way
- Low per-request cost and latency as reported in the README
- No generated text; all copy is human-written
Caveats
- A game/demo rather than a reusable library
- Requires hosted Jev access
Getting started
Play it at https://component-charades.vercel.app.At a glance
- Type
- Project
- Status
- prototype
- Language
- TypeScript
- Runtime
- browser
- Has code
- yes
Use cases and integrations
How curated lists describe it
A Taboo-style parlour game for design systems, refereed by Jev (TypeSafe System One model).
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide in the game?
Which of the library's components your description refers to, with a probability for every one.
When does it buzz?
When the top component reaches 0.85 confidence.
Is there a non-game use?
Find a component mode maps a job description to a component and says whether one exists.
Related projects
Source for browser chess and Chess960 where code computes each legal move's facts and Jev picks one per turn as a typed Choice.
ProjectproductionTypeScriptBrowser 1v1 sniper arena where Jev reads match state about 9 times a second and chooses movement, aim, ADS, firing and jumping.
ServicebetaTypeScriptInstrumented 2048 web lab where every move is a Jev Choice over four directions, with probabilities, latency and cost shown live.
ProjectbetaTypeScriptFifteen live arcade games where Jev and LLMs play each other, led by WikiRace in which Jev picks only real links.
ProjectbetaTypeScriptReal-time bot battles where you write a fighter's strategy in plain English and a System One model pilots it several times a second.
ProjectbetaTypeScriptBrowser extension that lets Jev, local Laya or an OpenAI-compatible model take over a running ra2web (WannaFire) RTS match.
ProjectbetaTypeScript