110 curated Jev entries. Official, runnable Jev recipes that show what decision models are good at, each with its own datasets, latencies, and numbers.
By format
110
Docs23
Repos & sites53
Videos8
Posts10
Playgrounds16
Updated
All 110 Jev Cookbooks
Search the full category, or filter by format and tag.
Fifteen runnable OpenRouter recipes for Jev covering support triage, database indexing, PII scanning, moderation, lead scoring, a browser agent, and a Gmail labeler, each with a small labelled dataset and saved live results.
A demo application that navigates a Neo4j graph one hop at a time: Jev gets the outgoing relationships as a Choice plus a goal Noul in one system_one call, and beam search follows the probabilities.
A ManiSkill demo where plain-English goals drive a simulated Franka Panda: Jev picks one primitive per step from a catalog of about 30 hardcoded skills, and Python executes the move. No training or policy.
Two WebAssembly builds of the Chocolate Doom 3.1.1 engine play the same Freedoom map from the same initial state, with a Jev Choice decision picking each tactical macro and a local controller turning it into inputs.
A Chinese and English lateral-thinking puzzle site hosted by Jev: Choice questions classify player questions and check proposed solutions against required facts, while code enforces confidence thresholds before a puzzle is marked solved.
A tldraw canvas controlled by voice and a webcam-tracked fingertip: every partial transcript triggers one Jev request through OpenRouter's Decisions API that decides the action, shape, colour, and place.
An original Civilization II browser harness for Jev: named Choice vectors cover empire policy, cities, research, diplomacy, units, exploration, and warfare, with live decision probabilities in a Roman-themed spectator display.
A social network where 10,000 computed personas read your post: Jev scores who the text is for to pick the first 600 readers, then answers one Choice per persona for its reaction.
A desktop Snake experiment where Jev assesses the board each tick and Python combines its answers into a move; one recorded run reached 29 foods in 461 ticks before trapping itself.
A terminal Wikipedia link-race tool: Chromium fetches article pages, Jev ranks their outgoing links by expected remaining hops, and Python runs a beam search without a search API or a controller model.
A self-hosted AI trading platform that can put Jev in front of strategy and Quick Trade entry orders as an auditable PASS/REJECT gate, with LLM fallback and deterministic bypasses for exits and protective orders.
A MuJoCo and Franka Panda workbench where Jev chooses the next manipulation step for pick-and-place, stacking, and obstacle-carry tasks, with rule baselines and side-by-side model comparison.
Browser lab where every 2048 move is one Jev Choice over the four directions with no heuristic fallback, surfacing the probability distribution, confidence, latency and token cost live.
Cloudflare Worker and React app that sends pasted corporate prose through about 45 typed Jev questions, then emits rustc-style diagnostics for passive-aggression, urgency and information density.
An experimental iOS app that turns an on-device Qwen3-VL model into a multiple-choice decision tool: it reads next-token logits for A/B/C and scores 2-26 supplied options without generating any text.
Two-stage Jev control of a Franka Panda in MuJoCo: Jev selects a task intent, then X/Y/Z directions and a gripper command executed against real contacts, with 43 of 50 Jev episodes succeeding versus 48 of 50 for a rule baseline.
An SO-101 workbench where an evaluation model or rules baseline picks bounded joint steps from typed candidate actions: decision runs are spend-capped via the Vercel AI Gateway and drive a real LeRobot arm or mock hardware.
A local, read-only Gmail triage app that sorts mail into five trays with Jev: each message becomes one experimental_evaluate call answering tray, urgency, and human-written questions, and 1,000 emails sort in about a minute for around 3 cents.
A joke npm package that left-pads a string by asking Jev to choose among space_0 through space_10 in one Choice, spending a model call where padStart() would do.
A side-by-side T-Rex game where local Laya runs on Apple Silicon through MLX and hosted Jev uses the TypeSafe API, showing choices, latency, survival streaks, and crash replays on one course.
A multiplayer demo workflow builder for Jev built with Liveblocks: wire Jev and LLM steps together, call the workflow through a REST API, and preview test runs with shared state and cursors.
A chat-shaped command bar where Jev picks the tool, arguments, confirmation, and reply type each turn, and code calls the MCP servers, so no LLM ever writes the reply text.
A fine-grained robot control study loop for LIBERO tasks: Jev layers an intent, a contact or motion family, and one of 27 control inputs, while reversible MuJoCo previews evaluate each candidate before execution.
A local 3D reference gallery where Jev chooses search phrases and highlights matching Met, NASA, Cosmos, and Prelinger Archive items from their text metadata.
Interactive browser experiments that expose Jev's judgments: a support-message Decision Lab and a driving simulation where structured sensor state becomes typed steer, brake, and overtake decisions.
An observable stealth game where Jev supplies guards' split-second judgments while deterministic code owns the 30 Hz simulation, validates every proposal, and exposes each decision in a Decision Lens.
An OpenRouter Labs recipe that has Claude rewrite a pasted prompt into Jev questions, then runs Jev on every row while a chat model races the same job, comparing answers, requests, time, cost, and accuracy.
A single shared chess game where visitors play against Jev, which weighs every legal move at once as one Choice question and shades the board with its probabilities.
A Telegram channel X-ray that pulls up to 1,500 public posts from the past year and has Jev rate each one, tallying news, ads, jokes, and clickbait into a shareable card.
A free no-signup demo that turns any question into Yes, No, or Maybe by calling jev-latest through TypeSafe's System One API, with live web search for questions that depend on current facts.
A hosted checker and public API that takes a pasted software idea and returns a Jev verdict with success potential, backed by a task guide and a POST /api/fit endpoint.
A live Val Town demo that re-asks 16 typed judgments about your text in a single API call on every keystroke, covering traits like passive-aggressive, AI-written, tone, and urgency.
A semantic mapping playground where you choose any two concepts as axes, add words or phrases, and Jev places each one by how strongly it relates to both axes.
Semantic search over GitHub's Terms of Service: one request ranks all 218 lines with a Choice while a Noul checks whether the document contains an answer at all, including when it should say no.
A two-stage extraction cascade: a mini model extracts, a Noul battery verifies each field in one request, and a 0.7 gate escalates to a reasoning model, sitting on the cost/quality frontier.
Turn natural language into typed function calls: closed-set arguments become Choice questions, optional ones become Noul questions, and 54 questions ship in one request for a trading assistant.
Decide which of 450 candidate product pairs from two beer catalogues describe the same entity: one three-level Score per pair splits 40 merges, 50 curator reviews, and 360 non-matches.
Catch wrong and hallucinated citations against RFC 7519 with a string match plus one Choice question: four accurate citations verified at 0.93 or more and all four planted failures caught.
A repeatability study: an 8-question moderation rubric run 15 times per condition shows a mean probability standard deviation of 0.0098, and a 0.60 uncertainty gate lifts agreement to 99.2%.
A 14-Noul claims-triage rubric over one insurance claim, repeated 15 times: mean probability standard deviation of 0.0102 at 111 ms per call, with a 0.30 to 0.70 band for human review.
Extract dates by asking seven Choice questions about a date's shape and parts, then resolving them in code: five of six examples auto-accepted and missing dates flagged at 0.46 confidence.
Score retrieved passages with four Noul questions each, then route in code: prompt injections dropped, false premises kept as conflicts, and evidence assembled separately for the generator.
Rank 182 agent skills in one request and re-read the top three in a second: over 488 requests, wrong skill loads fell from 16.8% to 7.3% and needless loads from 9.8% to 4.0%.
Reconstruct Markdown from unformatted text in two requests: 16 Noul line-pair questions stitch split sentences, then 62 questions classify each block's type and companions.
Turn wine tasting notes into 67 numeric columns for a CatBoost regressor: five rounds of reading its own errors cut held-out RMSE from 3.09 to 1.77 across 800 unseen reviews.
Regex finds candidate emails, phone numbers, and amounts; Jev picks the requested span so code copies a verbatim value that cannot be invented, with no free-text extraction required.
An experimental controller that has Jev choose NES inputs to play Super Mario Bros. from structured emulator telemetry instead of screenshots, logging every decision with probabilities and latency.
Retrieval is a sequence of bounded judgments. How a decision model slots into passage filtering, semantic find, reranking, and citation checks, and what stays in deterministic code.
Jev bills input tokens and gives output away, and questions in one request run in parallel. The design consequence is simple: ask more per call, then decide in code.
Jev Cookbooks#batching#patterns#search
Read article
From the Jev community
Watch walkthroughs, read the threads, and see what builders are posting.
YouTube: Retriever AI tests Jev on real browser tasks: where it worked, where it struggled, and what comes next.
YouTube: One terminal, three subscriptions, and Jev picking the model for Claude Code, Codex, and Cursor.
YouTube: A podcast segment on TypeSafe AI, its $40M raise, and the Jevons paradox behind the name.
Hacker News: Software Bombe and live Enigma demo where Jev ranks cribs and judges plaintext.