THE HUNDRED EYES
Interactive art piece where Jev reads a post as 100 fictional observers, picking an intuition and scoring four feelings for each.
Overview
THE HUNDRED EYES is an interactive media art work, documented in Japanese with an English README, that shows how one post is received by 100 fictional people, rendered as eyes in a dark 3D space. When the user posts, a blink travels from near eyes to far ones, four fixed observers with outsized reach surface first as the loudest voices, and a post-analytics view then counts all 100 equally across six intuition types.
For each fictional person, Jev answers one Choice for their intuition and four Scores: interest, liking, discomfort and desire to speak out. That is 500 questions per post, sent as four parallel requests of 25 people; results are applied only once all 100 valid answers arrive. The author defines the personas, short lines and voice reach, so Jev does not generate dialogue, and model confidence is not used as emotional intensity. Without a key the piece runs on simple local rules. It is built with Next.js, React, Three.js and Web Audio.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What is this observer's intuitive reaction to the post (six types)? | — |
| Score | How interested, favourable, uncomfortable and willing to speak out is this observer? | — |
How it works
- User submits a post.
- For each of 100 personas, ask Jev one Choice and four Scores, in four parallel batches.
- Apply results once all answers are valid; keep previous results on failure.
- Animate eyes, surface four loud voices, then show equal-count analytics.
Strengths
- Shows Jev driving visuals and sound rather than text.
- Falls back to local rules without a key.
- End-to-end tests run on synthetic responses.
Caveats
- Fictional personas; not real audience statistics.
- Sends the posted text to TypeSafe when a key is set.
Getting started
npm ci && npm run dev -- --port 3001; set TYPESAFE_API_KEY in .env.local to use Jev.At a glance
- Type
- Project
- Status
- beta
- Language
- TypeScript
- Runtime
- Next.js web app
- Has code
- yes
- Has tests
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Interactive media art: asks Jev one `Choice` and four `Score` questions per fictional observer to animate 100 eyes from a shared post, revealing four amplified voices before equal-count analytics expose the full distribution of reactions.
yibie/awesome-jev
Frequently asked questions
What does Jev decide?
For each of 100 fictional observers, an intuitive reaction type and four emotion scores about the post.
Does Jev write the observers' lines?
No, the author defines personas and lines; Jev only returns typed judgments.
Does it work without a key?
Yes, it uses simple on-device rules as a placeholder.
Related projects
Emoji autocomplete that asks Jev eight parallel questions per typing pause: emoji choice, emotion, mood, urgency, sarcasm and more.
ServiceprototypeTypeScriptCompanion that answers only with facial expressions: Jev picks one of 16 moods for each message and a blob avatar morphs its face.
ProjectprototypeTypeScriptWeb demo where you paste text and Jev suggests the best-fitting emoji.
DemoprototypeUndocumented app that uses Jev to classify text and pick an emoji reaction.
ProjectprototypePythonVisualizer that sends every passage of a novel to Jev for nine emotion scores plus intensity and draws the book as an aurora poster.
DemoprototypeTypeScriptDesktop app that cuts long videos into captioned short clips, with Jev reviewing candidate cuts for strength, context and completeness.
Appproduction