Jev directory
CommunityJev Demos & Experiments
Emoji Sift
Browser and terminal demo that ranks emoji against a free-text query by sending one Jev score question per emoji, then filtering by a tunable score floor and limit. Includes a bench harness over saved queries.
- Stars
- 0
- Reading time
- 2 min
- Published by
- Community
- Added
- Oct 2026
Highlights
- Browser and terminal demo that ranks emoji against a free-text query by sending one Jev score question per emoji.
- Answers land on a four-level scale from "unrelated" to "an obvious answer"; code ranks by score, drops below a floor, takes the top few.
- Two tuning knobs on the 0-3 scale: SIFT_FLOOR defaults to 1.5 and SIFT_LIMIT defaults to 12.
- Requires Node.js 22.18 or later and an AI Gateway API key; the key stays server-side.
- Includes a bench harness that runs every saved query and prints the top hit plus the first answer below the floor.
Questions about this entry
Short answers drawn from the linked project or page.
It sets the minimum score on the 0-3 scale for an emoji to appear at all. The default is 1.5; the docs note that lowering it to 1.2 roughly doubles results for goal-shaped queries while leaving strong descriptive queries untouched.
SIFT_LIMIT is the maximum number of emoji shown at once. It defaults to 12 and is set as an environment variable alongside SIFT_FLOOR.
The query is sent once as shared state, and each score question names an emoji and refers to the query in state rather than repeating it, so payload size stays constant. The API key stays server-side; the browser posts a query and gets back the ranked row.
Yes. pnpm run sift "things you can wear" runs one query with scores, pnpm run sift alone opens an interactive prompt, and pnpm run bench runs every saved query at once.
Jev has a single provider and no fallback; five consecutive runs of one query returned success, 503, success, 503. The server retries transient failures, while the terminal tools fail fast on purpose so endpoint trouble is visible.
A fixed yes/no bar returned 27 matches for "things that make noise" and 2 for "i need to lose weight". Ranking asks which emoji answer best, which is what makes open-ended queries work.
More Jev resources like this
Spanish-language prototype that uses Jev to classify event sign-ups and decide program, action, and message before sending, with a guardrail that approves, flags, or blocks outgoing text. Includes synthetic datasets, evaluation scripts, and a local engine that runs without a key.
Jev Demos & Experiments#madewithjev#demo#python
Communityblockstars-labs/30x-JEV
Privacy-first Next.js app that sorts saved X bookmarks into Top picks, Needs human look, and Skip using one Jev call per bookmark. Composite scoring, confidence-gated routing, and a deterministic mock mode that runs without an API key.
Jev Demos & Experiments#demo#typescript#routing
CommunityPirozshki/jev-bookmark-triage
A demo that has three decision models read recent news and price Polymarket and Kalshi questions, then compares their probabilities with live market odds and replays how each model changed its mind article by article.
Jev Demos & Experiments#community#github#demo
Community
yorkeccak/model-vs-marketTypeScript
yorkeccak/model-vs-marketTypeScriptThree TypeScript demos built on one pattern for students: Jev answers a typed question with a probability, your code decides what to do, and Claude works only where it is needed. Demos cover Jira triage, pre-flight checks and PR routing.
Jev Demos & Experiments#community#github#demo
Community
soyjorgepilo/jev-demoTypeScript
soyjorgepilo/jev-demoTypeScriptA browser visual IDE for composing Jev API requests: build a state plus Choice, Score, Noul and Box questions with image assets, preview and validate the request JSON, send it, and read the answers as visual cards instead of raw JSON.
Jev Demos & Experiments#community#github#typescript
Community
MoeclubM/PlayJevTypeScript
MoeclubM/PlayJevTypeScriptMinecraft survival bot built on Mineflayer that asks Jev four typed questions every 600ms — threat level, action, whether to eat, flee direction — plus a tactical goal-selection loop, with simulated emotions and persistent memory feeding each request.
Jev Demos & Experiments#demo#games#madewithjev
Communityakash-kamat/jev-craft
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.