Jev Classification

A logistics exception gate that reads carrier codes and free-text driver notes in one scan event and decides whether a shipment needs a human, which exception class it is, and how severe the delay looks.

A contextual brand-suitability scorer that judges each feed item against an advertiser's own versioned policy, returning a category, an exclusion flag, and a confidence for routing borderline items to review.

A podcast chaptering tool that labels transcript windows and marks boundaries as separate local judgments, classifying content, sponsor, and housekeeping segments without generating prose for the titles.

A zero-shot text classification service over plain HTTP: send text and a list of labels, get the best-fitting label and a calibrated probability per label, with Jev as the default and fast tier.

Cloudflare's first open decision models on Workers AI: Clef, a 27B model that reads text, JSON, images, or video, and Clef-flash, a 9B model for the latency-sensitive hop, each returning a probability per option.

The Hugging Face hub for Laya's open System 1 checkpoints: the 421M English model at the repo root, a 322M multilingual build in a subfolder, and the typed-decisions fine-tune, all Apache-2.0.

Supersonic Labs' 144.3M-parameter decision model on mmBERT-small: choice, score, and noul answers from supplied options in one pass, 73.15% on 2,000 typed decisions, Apache-2.0.

Calibre plugin that asks Jev typed questions about each book's genre and subjects, applies configurable per-category probability thresholds, preserves existing tags and leaves uncertain results for manual review. Supports single-book, selected-books and whole-library classification with caching, CSV export and restore.

Head-to-head benchmark of fast typed classifiers, TypeSafe Jev, OpenAI Decisions and Cloudflare Clef, against Claude models on a frozen, labeled corpus of 2,321 cases across 31 tasks, measuring accuracy, calibration, p95/p99 latency and registry-priced cost.

CLI and Python library that runs declarative YAML judgment tasks - classify, route, score, verify - natively on Jev Choice/Score/Noul, on any OpenAI-compatible backend, or on local keyword rules, with confidence gates, fallbacks and per-decision cost records.

A qualitative-research use case thread: classifying free-text participant responses without training a custom model first.

Classifies and splits PDFs, DOCX, and PPTX with local page text from LiteParse and category decisions from Jev.

An IMAP client that turns each inbox category into a Noul, then tags, moves, flags, or calls a webhook when the probability clears that category's threshold.

Sam Reghenzi's product-taxonomy benchmark: a Jev Choice at every level versus a gpt-5.2 agent that descends the tree and can be sent back by a judge.

Automate with Marc on using Jev to sort, tag, or score fixed answers, including a demo that tagged competitor ads.

OpenRouter's Banking77 run: 3,080 support utterances, 77 intents, Jev 1.13 against Claude Opus 5 on the same one-line criteria, with latency, cost, and a confidence cascade.

WitWireAI recreates a spreadsheet that asks Jev to score customer messages by meaning, and labels the clip as a demo rather than a benchmark.

A video demo of Jev working through years of hoarded Twitter bookmarks in one pass instead of reading them by hand.

Alan's team built an intern-review console on Jev after 4,000+ applications: it checked 3,518 applications in 4 minutes 54 seconds and surfaced 20 for humans.

An internal benchmark report: Jev zero-shot came within about 5 points of recall of a fine-tuned model at matched precision, at roughly $70 a month.

Ground Truth, a browser extension that shows the political framing, article type, topic, and loaded language of the news article you are reading.

WXT Chrome and Firefox extension that has Jev classify nonessential page elements through TypeSafe or Vercel AI Gateway, then stores reusable local hiding rules by page template. Paid analysis is manual by default, with an optional on-visit mode.

Experimental Hono router that matches an incoming HTTP request to a plain-language route description: one yes/no Jev question per description is evaluated in a single call, and the first registered route above the threshold wins.

Python MCP server that exposes classify, score, check, match, and screen tools to MCP-compatible agents, returning typed judgments over a closed answer set with token usage and latency. Ships generic triage and routing question examples.

A reproduction of Jev that turns any Qwen checkpoint into a decision model serving the same /v1/systemone schema (Choice, Score, Noul), with no training and no generated answer text.

An experimental Sutro CLI that evaluates CSV, Parquet, and JSONL data with Jev, asks you to label ambiguous and random audit rows, and uses GEPA to propose improved function definitions.

A Manifest V3 Chrome extension that finds ad-shaped DOM elements and asks Jev, one noul question per candidate, whether each is a paid advertisement; hits above the threshold are removed.

An Elixir client that treats Jev as a peer GenServer: you reply with questions and its typed answers arrive as messages you pattern-match, with network-free tests.

A Rust CLI and PyO3 library that streams JSONL and Parquet rows through Jev Noul and Score rubrics, writing clean and rejected records to disk without rewriting content.

An agent skill for deciding where typed judgments belong beside code, policy, and generation, with Jev examples for question design and abstention plus an offline probability and threshold evaluator.

A TypeScript library that walks a taxonomy and calls Jev once per level, so a choice can cover thousands of leaves despite Jev's 255-option limit per question.

An exploratory study of zero-shot ham, spam, and phishing classification with Jev: enriched email context raised main-set accuracy from 93.62% to 98.64%, within 0.23 points of a trained TF-IDF baseline.

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.

An async LangGraph workflow that sends a mocked email to Jev, receives a typed Choice between invoice and general, and routes it to a demo handler, with confidence and both label probabilities printed for inspection.

An independent local alternative to Jev: it takes English or Chinese context, questions, and candidate answers, and returns probabilities over the supplied candidates from a frozen Qwen3-4B backend.

A semantic code search CLI that asks one yes/no Jev question per function and ranks the returned probabilities, screening a whole codebase for a pattern in seconds for cents.

A Rust CLI that sends terse YAML or JSON question sets to Jev, validates them offline with 13 lint rules before any billed call, and prints parsed answers as JSON.

An async Python client for Jev with typed Noul, Choice, and Score helpers: questions batch into one request, and callers get probabilities, choices, and weighted scores instead of text to parse.

A RubyLLM 2 provider that adds Jev through structured output: build Noul, Choice, and Score questions with a schema builder, pass them to chat.with_schema, and read typed answers from response.parsed.

A community Rust client for Jev's System One API whose environment variables, defaults, retries, identification headers, and error kinds match the official Python and TypeScript SDKs.

A Rust CLI that classifies Git commit diffs and messages with Jev into bug fixes, security fixes with CWEs, and change types, then writes HTML or CSV reports from saved scans.

A local server and Python package that runs open models with Hugging Face Transformers or PyTorch and returns Choice, Score, and Noul answers read from next-token logits instead of generated JSON.

A grep-style CLI where the pattern is a plain-English description: each line becomes one Jev Noul question, judged concurrently with matches printed in input order, working on tail -f as well as files.

A local screening workbench for a folder of CVs: it asks Jev a small fixed set of typed questions about each candidate, scores the answers with plain arithmetic you can read, and shows a sortable shortlist.

A Rust derive layer for typed System One decisions: enums and structs become Choice, Score, and Noul questions with compile-time-checked, confidence-gated answers, plus a scriptable fake client for network-free tests.

An open-source, commercially hosted generative-engine-optimization platform that uses Jev judgments within a broader application to track brand mentions and placement in AI answers across ChatGPT, Claude, Gemini, and Perplexity.

An open, non-autoregressive System One model with published weights and a TypeSafe-compatible /v1/systemone API; the 395M-parameter OptionMarker checkpoint scores discrete and ordinal questions locally.

SQLite C extension that exposes Jev as SQL functions (jev_noul, jev_choice, jev_score) so queries can classify states such as ticket bodies directly, with only libcurl as a dependency.

Text-only classifier for 261 federal tax forms that sends each PDF page's text to Jev, which returns a probability over forms and page kinds behind a confidence gate; no model is trained.

Open multilingual System 1 decision models with published checkpoints for choice, score and noul questions, plus a router that dispatches each request to the right checkpoint in one forward pass.

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 Python library that adds .jev accessors to pandas and Polars: ask a natural-language question per row and get labels, scores, and full probability distributions back.

An open, local-first alternative to Jev that reproduces the Choice, Score, and Noul interface on commodity zero-shot NLI models with temperature scaling and conformal abstention, and runs offline with no API key.

A Next.js app that fetches a public GitHub pull request's dossier and runs one Jev systemOne judgment, returning risk scores, noul probabilities, review depth, and an approve, request-changes, or block verdict.

Pick one option from a set you define. Returns the option, the full probability distribution, and 0-1 confidence, with up to 255 options and parallel questions that barely add latency.

The yes/no primitive: one probability from 0 to 1, where the value is the answer and no separate confidence is needed. True and false criteria pin down subtle boundaries.

Use the answer to decide what to do and confidence to decide whether to act: the voice-banking example routes below 0.6 to a human and needs 0.85 or more to auto-approve a transfer.

Ask everything the system might need in one request, then let code throw away what it does not use. The ticket-triage example sends a category plus four speculative questions in parallel.

Classify intent with a Choice and complexity with a Score, then send each branch to deterministic code, a specialist model, or a person, with a 0.5 intent-confidence floor.

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.

A pre-alpha PostgreSQL extension for categorical work in SQL: Noul, Choice, and Score functions plus batched variants that classified 38 strings in two requests.

vogel ran Jev over 1,500 of his own emails to test classification quality and posted the results as a video, calling it the most impressive model he has tried for the task.

Gojiberry's Romàn reports scoring 700 high-intent leads and personalized outreach messages in 40 seconds for $0.09, with confidence scores and lead-message mismatch detection.

Classify documents through deep patent, retail product, biomedical, and source-code hierarchies using parallel beam search over Choice probabilities.

Classify SEC annual reports into 75 industry groups with one Choice each, then read confidence to decide when to fall back to a broader division.