Use-case gallery

Open-source Jev-ify any HF model (SimpleJev)

Builder ships SimpleJev.ai: open-source library to turn any Hugging Face model into Jev-like Choice/Noul/Score API, including vision.

choice · noul · score

Fast context compaction scoring (ecosystem roundup)

72-hour ecosystem roundup highlighting fast-jev-compaction: Jev scores tool calls/results for relevance and drops unneeded context instead of LLM rewriting history.

score · noul

Open 151M typed-decision calibration vs Jev

Builder audits LocalLLaMA/typed-decisions harness bugs, then shows a CPU TF-IDF+logistic baseline (11s) beating Jev on ECE and matching customer-service accuracy; shares a reproducible GitHub breakdown.

choice · noul · score

BANKING77 classifier evaluation

Compared Jev on the BANKING77 intent dataset, reporting 92.40% accuracy versus 93.66% for a fine-tuned BERT model.

score

Open-source unkillable Mario (Jev + microsandbox)

Builder shares mario-never-dies: an unkillable character controlled with Jev decisions plus microsandbox.

choice

CLI and stdio MCP glue

Reports an unofficial Jev CLI plus stdio MCP with JSON output and uv install as easy-to-try integration glue.

choice · noul

Open-source Rust port of Jev SDKs

OpenAgents open-sources a Rust port of the Jev SDKs under their monorepo and asks TypeSafe for an official Rust SDK.

choice · noul · score

Open-source Jev Ultrafast browser agent

Browser Use\u2019s jev-ultrafast: each DOM snapshot becomes an indexed action space; Jev picks operation and target in one request (small LLM only for TYPE_TEXT). Public demo ~7s Google Flights for ~$0.0039.

choice

Jev Tweet Radar Chrome extension

A MIT-licensed Chrome extension uses Jev probabilities while browsing X to estimate whether a post is worth interacting with; it supports user-defined interaction types and Chinese/Japanese/English.

choice · noul

Expense categorization in an open-source app

Jev is wired into an open-source expense app to automatically attribute categories to new expenses; the post includes a mobile UI demo.

choice · noul

CUA-S1-FORMS specialist

CUA-S1-FORMS applies the System One pattern to forms: a 706,048-parameter, 2.8 MB MIT model chooses FILL/CHECK/CLICK/SKIP in one forward pass instead of making an LLM reason through the form.

choice · noul

Jev in MobileWorld benchmark

MobileWorld reports a 389 ms median Jev API latency and 27.4% task success, contrasted with a 77.8% end-to-end Qwen3.8-Max reference; full trajectories and integration are planned in the benchmark repo.

choice · noul · score

Realtime brain scanner

The realtime brain-scanner project reports 7/7 simple questions landing in its top three, 275 ms median response time, and about $0.00017 per call, while also publishing failures.

choice · noul

Stateful Jev / Jev One

A stateful Jev implementation is demonstrated as a JavaScript library for stateful applications, with a video prompt asking Jev to take care of a place and choose actions over time.

choice · noul · score

Jev practice speed harness

A public practice project lets developers obtain a Jev API key and test their own decision speed against Jev, turning the API into a repeatable latency practice harness.

choice · noul · score

NotHotDog multimodal decision app

The open-source nothotdog demo uses a vision model to write one sentence about an image and Jev to decide from that description; the author reports 1002 ms end-to-end versus 327 ms for the Jev half and displays both bars.

choice · noul

Verdict OSS-Jev benchmark update

The author reports that the fully open-source Verdict Jev beat both official TypeSafe Jev and Laya on their benchmark, following an earlier Verdict 1.0 score of 26.10% accuracy.

choice · noul · score

Curated Jev application catalog

Aster Dev points builders to a GitHub-curated list of Jev/System One applications, libraries, and resources for code-by-example exploration.

choice · noul · score

Laya open-source System One alternative

The post describes Laya as an open-source 421M non-autoregressive System One model with one-forward-pass probability decisions, no text generation, and local GPU execution, claiming faster and better-calibrated results than Jev.

choice · noul · score

CUA-S1 open-source release

Cua announced CUA-S1 as a family of small, specialized System One computer-use models and open-sourced CUA-S1-FORMS as its first release.

choice · noul · score

Remote browser agent loop (Stagehand + Jev)

Observe a page, send the a11y tree as state with actions as questions; Jev picks the next operation/targets, Stagehand executes\u2014~$0.001 per task at near-instant speed.

choice · noul · score

Block-speed buy/sell trading bot (Monad + Kuru)

Jev chooses buy or sell from an asset price feed each ~300ms Monad block and places post-only limits on Kuru\u2019s on-chain book.

choice

Real-time Doom bot (~10 QPS)

TypeSafe launch demo: Jev drives Doom at ~10 calls/sec (~$7/hour) for real-time game decisions from structured game state.

choice · noul · score

Wikipedia link race (high-cardinality Choice)

Wikiracing demo: among hundreds/thousands of outbound links, Choice picks the next page without hallucinating invented links.

choice

Bookkeeping outsourcing with constrained choices

Japanese post sharing an image summary of how Jev (TypeSafe) could change bookkeeping outsourcing via constrained decisioning (#Jev).

choice

Joint email triage decisions

Shows joint Jev choices for an email: billing vs security and urgent vs not urgent, with probabilities for combined outcomes.

choice

Rock-paper-scissors Choice/Score demo

Demonstrates Jev Noul, Choice, and Score primitives with a rock-paper-scissors example.

choice · noul · score

Multi-axis text classification

Tested Jev through OpenRouter for multi-axis text classification and compared its accuracy against another model.

choice · score

Noul probability-composition test

Probed Jev Noul with independent-event probabilities to verify whether combined probabilities match expected composition.

noul

Jev-compatible probability API demo

Built a demo that chooses from predefined answers, normalizes token probabilities into a distribution, and returns them through a Jev-compatible API.

choice · score

System One interface approximation

Approximates a System One interface using model logits, softmax, and logprobs to return typed probabilities.

score

Customer-support decisioning

Uses a support message as state and asks in parallel whether it is urgent, which team handles it, and how frustrated the customer is.

choice

Agent-driven analysis ranking

Feeds raw data and roughly 50 bundled questions to Jev so scores drive the next round of analysis until the important signals emerge.

score

Parallel support-ticket questions vs sequential LLM

Explains Jev Choice/Score/Noul and contrasts asking four support questions (urgent, escalate, department, refund) in one parallel call versus sequential ChatGPT answers; also sketches lead classification.

choice · noul · score

Sky-color playground decision

Playground test chose Baby Blue vs Electric Blue vs Gray with probabilities in milliseconds; connects the pattern to Doom, automation, routing, and scoring.

choice

Agent harness routing and AutoMode

Describes Jev changing agent harnesses through model routing and AutoMode classification of tool-use risk.

choice

Five high-leverage architecture patterns

Lists tool-call compaction, RAG chunk filtering, browser click selection, shell safety gates, and high-throughput log triage as Jev patterns.

choice · noul · score

Ad-library classification at scale (reported)

Reports a builder demo classifying 1,891 ads in 19 seconds for $0.12 with Jev, and breaks down related decision-model experiments in an X Article.

choice

Medical AI exploration with System One

Radiology/medical-AI lab notes that System One-style decisioning may fit medical workflows better than text generation, with ongoing verification linked in a Note writeup.

choice · noul

Getting started after connecting Jev to Codex

Builder shares an X Article on getting started with Jev after connecting it in a Codex-oriented workflow.

choice · noul · score

Buying-signal triage at million-mention scale

Explains a large Jev workflow that scores ~1.76M mentions for buying signals, ranks intent/product fit, and drafts replies\u2014separating marketing claims from the decision-model job.

choice · score

Agent-action evaluation with Jev

Reports Eve reworking LLM-as-judge support so Jev can define evaluations and validate agent actions.

noul · choice

Codemode and procedures store idea

Suggests building codemode plus a procedures store around Jev as a practical agent workflow.

choice · noul

Jev versus Siri command routing

Ran Jev against the 50 most-used macOS Siri commands and found median requests nearly 5x faster including tool calls.

choice

Selective OCR for chart pages

Describes processing only pages that need OCR for charts, avoiding OCR on other pages before classification.

choice · noul

Text-PDF classification gate

Notes Jev can classify text-only PDFs quickly, while PDFs with images/charts still require full OCR first.

choice

Browser next-click selection

Uses Jev to rank all visible anchors and choose the next click, with harness instructions steering it toward the goal.

choice

Cheap-flight scanning

Reports success using a Jev repo example to scan for cheap flights across a large time window.

choice

Multilingual family-assistant routing

Ran Jev on Loon family-assistant routing tests in English, French, and Arabic, achieving 98%, 100%, and 92% respectively at 164ms median.

choice

Typeahead UI prediction

Demoed Jev in a typeahead UI that predicts what to update before the user finishes typing.

choice

Qdrant decision-model integration

Plans to integrate Jev with Qdrant as a small, specific decision model.

choice

JevSQL angry-ticket predicate

Built a SQL predicate, jev_bool(body, 'Is this customer angry?'), for typed natural-language filtering of tickets without embeddings or an LLM writing SQL.

noul

Jev as agent/model router

For daily product, research, content, and ops agents, the proposed shipped pattern keeps an LLM for research/writing/coding while Jev chooses the agent, model, or escalation path.

choice · noul

Jev gate before risky tool calls

A concrete deployment pattern puts Jev before shell, deploy, and other irreversible tool calls, runs it in shadow mode for a week, then automates only the low-risk path rather than replacing the LLM.

choice · noul

Google Flights automation demo

A Japanese Jev article/demo shows an AI operating Google Flights and finding flights in about seven seconds at normal playback speed, alongside a 20-use-case/API guide.

choice · noul

24/7 HFT trading design

The author describes a Jev-based high-frequency trading design with calibrated buy/sell decisions under 100 ms and one decision per block, with a from-scratch implementation article.

choice · noul

Support inbox and driving simulation

In a week of hands-on experiments, the author used Jev with a support inbox and a driving simulation to show typed decisions and probabilities in practical application flows.

choice · noul

Jev drives the dre diagram editor

In a working demo, plain-English sentences build diagrams in the keyboard-first dre editor; Jev makes each decision in roughly 0.6 seconds and the app presses the corresponding keys.

choice · noul

Jev plays Manic Miner

A quick experiment wires Jev into the retro game Manic Miner and publishes a video showing the model selecting gameplay actions.

choice · noul · score

Classifier.dev over Jev for routing and RAG filters

A no-key HTTP service built underneath TypeSafe's Jev accepts text plus labels and returns a label and confidence for up to 1,000 items per call; the linked explainer demonstrates routing, RAG filtering, and log classification.

choice · noul

Inquiry-form routing

A live Japanese demo classifies and routes the contents of incoming form inquiries with Jev, showing the resulting workflow board at monbankun.hellocraftai.com.

choice · noul

Kalshi trading lab

A live Kalshi lab feeds BTC and ETH market data into Jev, applies code-enforced risk limits, and lets viewers watch 44 typed decisions choose, wait, and reassess; median response time is reported as 380 ms.

choice · noul

Bounded routing with cheap abstention

A concrete guardrail design uses Jev for bounded, high-confidence routing and hands cases whose rationale changes the action to an LLM; abstention keeps out-of-lane questions cheap.

choice · noul

Typed Jev API as a decision layer

The post highlights production-oriented Jev API capabilities beyond plain classification: yes/no questions with probabilities, numeric scoring, mixed question types in one request, and structured state input.

choice · noul · score

Open-source Jev Browser (MCP/CLI + Playwright)

jkudish/jev-browser: browser-use tool driving a real headless Chromium via MCP, CLI, or library, with Jev choosing click/type/select/stop. Cited on systemonemodels.org.

choice · noul · score

Open-source jev-chat command bar (no text generation)

w3cj/jev-chat: chat-shaped command bar that uses Jev Choice/Noul to select tools and never generates free text\u2014replies come from tool outputs only.

choice · noul

Open-source jev-experiments (~20 latency demos)

dabit3/jev-experiments: monorepo of ~20 self-contained latency-focused demos for routing, moderation, search, native macOS, and real-time control with Choice/Noul/Score.

choice · noul · score

Open-source jev-grug Choice word tournament chat

mkotlikov/jev-grug: toy cave-chat that makes Jev converse by repeatedly choosing the next word from a compact vocabulary (intentional inversion of System One design).

choice

Semantic search replace/supplement embeddings

Replace or supplement embeddings in RAG with semantic search, scoring, and ranking over candidates.

score · noul · choice

Query\u2013candidate relevance scoring

Score how well each retrieved candidate answers the query before generation.

noul · score

Pairwise result reranking

Rerank shortlists with pairwise comparisons using TypeSafe questions.

choice · noul

Cross-encode query and candidate

Cross-encode query+candidate for higher precision than bi-encoders alone.

noul · score

Select useful context for downstream AI

Pick which snippets are useful enough to pass into an LLM or agent step.

noul · choice

Systematic review paper screening

Screen papers against inclusion/exclusion criteria for systematic reviews.

choice · noul

Qualitative theme labeling

Label interview transcripts, surveys, and field notes with predefined themes.

choice

Claim\u2013citation support check (science)

Check whether cited passages support claims in manuscripts and summaries.

choice · noul

Flag missing methodological details

Flag missing controls, dataset descriptions, and experimental settings.

noul

Research knowledge-graph entity linking

Identify entities and relationships across papers linking findings to passages.

choice · noul

Custom LLM model router

Build a router that chooses which LLM receives each prompt using Jev.

choice · score

Intent and domain classification for routing

Classify intent and domain to set routing rules and thresholds.

choice

Estimate difficulty and risk of prompts

Estimate prompt difficulty/risk before escalating to expensive models.

score

Escalate hard requests to expensive models

Escalate only requests that need a more capable/expensive model.

noul · choice

Jailbreak and prompt-injection detection

Detect jailbreaks and prompt injection on LLM inputs at a fraction of LLM cost.

noul · score

Policy violation and sensitive-data checks

Identify policy violations and sensitive-data exposure on I/O.

noul

Tool-call error detection

Detect tool-call errors and response-quality failures in real time.

noul · score

Structured guardrail logging

Log structured check results and probabilities for harness failure tracing.

noul · score

Semantic code linting in CI

Add semantic lints for team coding conventions and writing guidelines in CI.

noul · choice

Probabilistic NL features for ML

Extract probabilistic features from natural language for predictive models.

noul · score

Autoresearch feature discovery loop

Propose feature definitions and evaluate predictive value on held-out truth.

noul · score

Resume and interview criteria scoring

Evaluate resumes/applications/interview feedback against job criteria.

score · noul

Candidate\u2013role matching and routing

Match candidates to roles and route to hiring managers; escalate uncertain cases.

choice · noul

ICP lead matching

Match company profiles and inbound messages to an ideal customer profile.

noul · score

Buyer intent and pain-point detection

Detect buyer relevance, pain points, and purchase intent; prioritize leads.

noul · score

Support ticket multi-axis classify

Classify tickets by issue, product area, and customer intent.

choice

Call transcript issue/commitment extraction

Extract issues, commitments, and follow-ups from call transcripts.

choice · noul

Urgency, frustration, churn, refund signals

Detect urgency, frustration, churn risk, and refund requests on tickets.

noul · score

Support queue and automation routing

Route cases to team, queue, or automated workflow.

choice

Verify support replies vs policy

Verify support responses against policies and the customer request.

noul · choice

Insurance FNOL classification

Classify first-notice-of-loss, adjuster notes, and supporting docs.

choice

Claim complexity, missing info, fraud flags

Detect claim complexity, missing information, and potential fraud indicators.

noul · score

Straight-through vs specialist claim routing

Prioritize claims for STP or specialist review; escalate uncertain cases.

choice · noul

AML/KYC suspicious narrative scoring

Evaluate transaction narratives, KYC docs, and alert histories for suspicion.

noul · score

Entity match across inconsistent names

Match entities across inconsistent names, profiles, and records.

score · choice

Financial alert prioritization

Prioritize alerts by risk, relevance, and evidence quality.

score

Contract and filing classification

Classify contracts, policies, regulatory filings, and marketing claims.

choice

Missing clauses and prohibited claims

Detect missing clauses, prohibited claims, and policy violations.

noul

Legal requirement verification

Verify documents against explicit legal or compliance requirements.

noul · choice

Product listing normalize/classify

Classify and normalize product listings across inconsistent seller catalogs.

choice

Product attribute extraction

Extract product attributes from titles and descriptions.

choice · noul

Counterfeit, review abuse, prohibited listings

Detect prohibited listings, counterfeit signals, review abuse, policy violations.

noul · score

Nuanced community moderation

Apply company-specific criteria to decide which posts meet moderation standards.

choice · score

Toxicity, spam, PII, opt-out detection

Detect toxicity, harassment, spam, fraud, unsafe advice, PII, opt-outs.

noul · score

Severity\u00d7confidence allow/warn/block

Combine severity and confidence to allow, warn, review, or block.

score · noul

Ad brand safety and suitability

Classify brand safety and audience suitability for creatives and placements.

choice · score

Ad regulatory claim compliance

Check regulatory compliance and prohibited claims in ads.

noul

Creative quality and LP alignment

Evaluate creative quality and ad-to-landing-page alignment.

score · noul

Game chat abuse moderation

Moderate in-game chat; detect abuse, toxicity, suspicious behavior.

noul · score

Player frustration and engagement scoring

Annotate content and score frustration or engagement from reports/reviews.

score

Player churn and support routing

Detect churn signals and route player-support requests.

noul · choice

Incident-to-risk probabilistic indicators

Convert incident reports and vendor assessments into probabilistic risk indicators.

noul · score

Underwriting risk type classification

Classify risk types and detect suspicious characteristics for underwriting.

choice · noul

Demand forecasting semantic enrichment

Enrich forecasts with purchase intent, urgency, and product interest from text.

noul · score

Supply/competitive pressure themes

Detect supply concerns, competitive pressure, and emerging demand themes.

noul

KG relationship and entity typing

Classify relationships and entity types when annotating knowledge graphs.

choice

Contradiction detection in records

Detect contradictions between records or claims in a knowledge graph.

noul

Self-consistency Noul review gating

Route uncertain Noul probabilities to human review while keeping underlying values visible.

noul

Self-consistency Choice moderation

Add uncertain outcome to moderation; compare label agreement with share of automatic actions.

choice

Parallel regulatory briefing questions

Batch many questions (e.g. 13 GDPR briefing Qs) in one call\u2014cheaper and faster with same answers.

choice · noul · score

Legal BM25 + TypeSafe rerank (CLERC)

Rerank BM25 shortlists with one question per query\u2013candidate pair to lift legal retrieval accuracy.

noul

Line-by-line semantic document search

Score ToS line IDs against a plain-language query; Noul checks whether an answer exists.

choice · noul

Recover Markdown structure from plain text

Stitch hard-wrapped lines and classify blocks (heading, list, code, callout) with companion questions.

choice · noul

Map NL trading requests to typed functions

Map function names and closed-set args to confidence-aware questions for ordinary typed functions.

choice · noul

Agent skill suggestion from large catalog

Rank Hermes catalog skills and decide whether a turn needs a skill at all.

choice · noul · score

Beer catalogue entity alignment

Score candidate pairs as merge / leave unlinked / hand to curator across two catalogues.

score

Classify RAG passages for injection/contradiction

Score retrieved passages; keep/flag contradictions; drop hidden instructions or injections.

noul · choice

Double-check citations against source

Choice whether quote context supports the claim; confidence flags for human review.

choice

LLM I/O hazard guardrails battery

Noul hazards + severity Score on every LLM input/output; threshold to pass/review/block/support.

noul · score

SDE cascade mini\u2192verify\u2192reason

Two-stage structured-data-extraction cascade for big-model quality at lower cost.

noul · choice

Date part extraction then code resolve

Ask TypeSafe for named date parts, resolve/validate in code with confidence-based review.

choice · noul

Regex candidates + TypeSafe span pick

Regex finds emails/phones/amounts; TypeSafe selects the requested span for verbatim normalize.

choice

Hierarchical beam-search classification

Classify through deep patent/retail/biomed/code hierarchies via beam search on Choice probs.

choice

Autoresearch \u2192 CatBoost features

Autoresearch loop proposes TypeSafe questions as numeric features for a CatBoost regressor.

noul · score

SEC report industry Classification + confidence

Classify SEC annual reports into industry groups; use confidence to fall back to broader division.

choice

Speculative fan-out many questions

Send many speculative questions in one call; code decides which answers matter.

choice · noul · score

Confidence-gated act vs escalate

Use answer for what and confidence for whether to act or escalate.

choice · noul · score

Composite weighted scoring

Break complex judgment into atomic scores; combine with code-owned weights.

score

Intent routing to code/LLM/human

Classify requests and route to deterministic logic, specialist LLM, or human.

choice

Smart-home request evaluation demo

Demo: evaluate smart-home user requests with typed System One questions.

choice · noul

OCR screen \u2192 next computer action

Community typesafe-computer-use: OCR the screen then classify the next desktop/computer action (~$0.0002/step).

choice · noul

Emulator Mario from structured state

Community typesafe-mario: play Super Mario Bros from structured emulator state (same shape as Doom demo).

choice · noul

Staged code-review workflow (jev-review)

Community jev-review: staged code-review workflow with local dashboard using typed review questions.

noul · score · choice

Software factory orchestration (foreman)

Community foreman: orchestration layer for a software factory making typed progress/quality decisions.

choice · noul

LangChain model-routing middleware

LangChain TypeSafe middleware: Jev assesses the user message and selects fast vs powerful model for the run.

choice

LangChain AutoMode dangerous-tool guard

AutoModeMiddleware uses Jev to classify risky tool calls (e.g. bash) and block before execution.

noul · choice

Email triage at scale (Ryan Vogel)

Public LangChain writeup cites Ryan Vogel doing email triage at scale with Jev decisions.

choice · noul · score

Support department routing (Cloudflare Workers AI)

Cloudflare docs example: route support requests to billing/tech/other with Choice on login/payout failures.

choice

Structured refund review (Cloudflare)

Evaluate refund requests against order+policy with Noul/Choice/Score in Workers AI example.

noul · choice · score

Account risk escalation (Cloudflare)

Score account risk and Noul whether activity warrants immediate escalation.

score · noul

Newsletter sponsor inquiry filter

Flavio Copes pattern: Noul for real sponsorship inquiry, Choice for category, Score for ask specificity.

noul · choice · score