Explore the catalogue ↗   First call   Adapt a project   Independent reports   中文 

_{Counts describe saved link and evidence records, not current CI passes or runtime tests. About these counts}

• Jev is a decision model from TypeSafe AI. It does not write text — you hand it state plus typed questions and it returns typed answers with calibrated confidence, fast and cheap enough to sit in an agent's inner loop.
• This repo indexes public examples of using it, organised by the decision being made. The resource you read this week is disposable; the decision pattern is not.
• How to assess it: every row names its source. Call-site citations, primitive claims and caveats are recorded where available, so you can inspect what was read and what remains untested.

Three primitives. Every pattern below is built out of them, and the asymmetry in the last row is the single most common source of bugs.

Input is text only — string, JSON object, or array of text. Context is 64k tokens per request, 32k for the state plus the longest question. Output tokens are free. There are no published weights, so it cannot be run locally. Full cross-platform differences: docs/compatibility.md.

Six things in reading order. Hand-picked, because "most starred" is not the same as "read this first".

1. Quickstart

   The canonical first call: one support ticket, one Choice, one Score and one Noul in a single request, in Python, JS and cURL.

2. Jev 1.13 known limitations

   The most useful page in the docs and the least linked. It explains, among other things, that a Choice over options and one Noul per option answer different questions.

3. Example: three primitives in one request

   Written from the official API reference and checked field by field against it, but not executed against the live API.

4. fast-jev-compaction

   Exactly two nouls per tool call: does knowing this call happened still matter, and is the full output still needed verbatim. Despite the word "scored" in its own description, no score primitive is used.

5. ai-cookbook: Jev track

   The best structured tutorial found. It states plainly that typed output does not guarantee a correct decision, lists the documented weaknesses, and qualifies its own cost illustration rather than selling it.

6. Hermes Agent: Jev compaction evaluation

   The single most credible row in this catalog. Recall came out below their existing summariser, and at a matched context budget it tied plain recency ordering. Cost was genuinely far lower. Publishing a negative result on a hyped model is rare.

Every decision pattern, sized by how many entries this catalogue contains. Use the pattern index below the chart to jump to a section. A zero is a research gap, not a rendering bug.

All 18 patterns have at least one catalogue entry. Coverage does not imply runtime testing or equal maturity. See docs/status.md.

Pattern index · jump to the examples

Independent measurement reports in the catalogue, including negative results that help explain where an approach fails. These are the original authors' measurements; this repository has not independently reproduced them. Check each report's dataset, method and model version before comparing results.

• Hermes Agent: Jev compaction evaluation
  Ported the Jev compaction approach, measured it against their shipping summariser, and published the conclusion not to adopt it.
  _{Benchmark · ★248,479 · Py · noul}

  The single most credible row in this catalog. Recall came out below their existing summariser, and at a matched context budget it tied plain recency ordering. Cost was genuinely far lower. Publishing a negative result on a hyped model is rare.

• worldmonitor: news threat classification
  Two Choice questions over threat level and category, held in shadow mode after a blind evaluation found Jev merely tied the incumbent model.
  _{Benchmark · ★87,298 · TS · choice}

  Caveats: shadow mode

  Wired in but deliberately inert: by their own statement nothing Jev returns reaches a label, a cache row or an alert. Ships a golden fixture. A model to copy for how to trial a new model without betting production on it.

• no-mistakes: Jev review pre-brief, measured and retired
  One Score per candidate file to pre-brief code review — measured twice, then removed: more billed input for essentially no wall-clock gain, and offline replay showed the candidate list could not reach where review findings land.
  _{Benchmark · ★8,617 · Go · score}

  Removed in PR #1165 (2026-09-22). Their offline measurement found the candidate generator excluded changed files by construction while nearly all review findings sit in changed files, and that per-file excerpts made the list less precise at higher token cost. The code is gone from the default branch, so this row cites the change that removed it.

• hippo-memory
  Biologically-inspired memory for AI agents. Decay, retrieval strengthening, consolidation. Zero runtime deps, SQLite, MCP. Benchmarked retrieval with an opt-in TypeSafe Jev reranker.
  _{Benchmark · ★756 · kitfunso · TS}

• Probing Jev's behaviour with repeated API calls
  Independent Korean-language notes reporting that reversing the order of options shifted a probability enough to flip a 0.9 threshold.
  _{Benchmark · ★190 · Py}

  Caveats: no licence · unverified claims

  The most actionable engineering caveat found anywhere: if option order alone can move a probability past your threshold, your threshold is not as stable as it looks. Independent and unreplicated, so treat the magnitude as indicative.

• jevbench
  JevBench v1 - a benchmark for Jev-class typed decision models: smart, cheap, fast, reliable, open.
  _{Benchmark · ★107 · fstandhartinger · Py}

• jev-arena
  An introduction to Jev with hands-on tests: Choice, Score and Noul turn natural language into typed judgements for classification, scoring and routing, compared with DeepSeek on comment labelling, speed and results, with CSV import, replay and offline reports.
  _{Benchmark · ★97 · nanmicoder · JS}

• windtunnel
  A WebMCP benchmark, measures WebMCP against other browser-agent interfaces.
  _{Benchmark · ★80 · nekuda-ai · TS}

• jev-robot-control
  Jev against two LLMs on direct Cartesian control of an xArm7 in MuJoCo — intent, movement and gripper each step — with recorded responses, trajectories and replays. One seed-0 trial per controller, not a success rate.
  _{Benchmark · ★44 · openroboto-ai · Py}

• typesafe-ai-benchmark
  A gateway that mimics the structured-output shape, used to benchmark against it.
  _{Benchmark · ★38 · iammrduncan · TS}

• jev-capability-atlas
  Independent, evidence-based map of when TypeSafe's Jev actually holds up vs. breaks down — real API-call receipts, not a leaderboard. 中文為主的雙語 repo。
  _{Benchmark · ★26 · zaious · Py}

• smartmoney-cub
  Read-only trading journal and review harness: Jev typed judgments, agent integration, and a reproducible finance benchmark. No orders, no advice.
  _{Benchmark · ★26 · myc0576 · Py}

• jev-benchmarks
  Probability-aware evaluation for typed decision models: calibration, selective risk, latency, and reproducible benchmarks.
  _{Benchmark · ★17 · abdelstark · Py}

• jev-rag-benchmark
  Reproducible benchmark for measuring Jev reranking quality, latency, and cost in RAG
  _{Benchmark · ★14 · erendikmenn · Py}

• pdf-race
  Docling → Jev vs Docling → Gemini 3.8 Flash vs Gemini reading the PDF: same documents, one clock, scored against arXiv's own metadata
  _{Benchmark · ★9 · goodrahstar · JS}

• jev-dspy-lab
  Reproducible calibration and selective-risk benchmarks for Jev/TypeSafe decisions in DSPy workflows
  _{Benchmark · ★7 · jmanhype · Py}

• jev-rerank-bench
  An independent head-to-head against dedicated rerankers across fourteen datasets.
  _{Benchmark · ★7 · anessbelbati · Py}

  An independent measurement rather than a vendor figure, and a direct comparison against purpose-built rerankers — the comparison that matters for the search-ranking pattern.

• jev-benchmark
  Benchmarks and a playground for TypeSafe's Jev (System One) model: chess, and who-is-the-player-talking-to for speech-to-text game NPCs
  _{Benchmark · ★6 · wondertwins · Py}

• jev-korean-benchmark
  Reproducible early-access evaluation of Jev on Korean understanding and medical text, with runtime and cost evidence
  _{Benchmark · ★6 · mahlernim · Py}

  Caveats: no licence

• jev-ood-calibration
  Independent calibration test of TypeSafe's Jev on a task it cannot have seen: 900 rule-generated support tickets (choice / score / boolean) plus 3 public benchmarks via Vercel AI Gateway. Raw responses, ECE with noise floor, temperature refit, per-type sign of miscalibration. Reproducible for ~
  _{Benchmark · ★6 · scienthoon · Py}

• jev-search-rerank-eval
  Does a TypeSafe Jev rerank beat embedding search? Graded relevance eval (9,831 pairs, 164 zh/en queries) over the Agent Skills Hub catalog, with the judge-circularity bias measured.
  _{Benchmark · ★6 · zhuyansen · Py}

• jev-code-review-benchmark
  Comparing Jev, Gemini Flash, and Claude Fable on Python code review rules: cost, speed, accuracy, and consistency. Includes results, charts, and reproducible experiments.
  _{Benchmark · ★5 · gemanor · Py}

• jev-little-airways
  A show-and-tell capability study for Jev, TypeSafe's System One decision model.
  _{Benchmark · ★5 · lbotinelly · TS}

• jev-phishing-bench
  Jev (TypeSafe) vs Claude Haiku 4.5 on 2 000 phishing emails: accuracy, calibration, latency, cost. Reproducible benchmark.
  _{Benchmark · ★5 · anisselbd · Py}

  Caveats: no licence

• legalforecastbench
  LegalForecast-MTD benchmark alpha and official evaluation workflows
  _{Benchmark · ★5 · johnhughes3 · Py}

• jevarena
  Open-source BYOK arena for Jev and other AI judges. Find failures, compare quality, cost, and latency.
  _{Benchmark · ★4 · chenmingtang830 · TS}

• sysone-bench
  First independent head-to-head benchmark of System One decision models (Laya vs Jev) on byte-identical inputs
  _{Benchmark · ★4 · instax-dutta · Py}

• ego-jev-ultrafast
  Jev drives your Ego Lite browser: one typed-choice request per step. Single-file, zero-dependency port of browser-use/jev-ultrafast with multi-model benchmarks and extra guardrails. Unofficial.
  _{Benchmark · ★3 · shikaizhong-design · JS}

• jev-does-not-play-dice
  Experiments on Jev’s probability calibration, uncertainty reporting, and forecast probability preservation.
  _{Benchmark · ★3 · kantahayashiai · JS}

• jev-exploration
  Jev (TypeSafe) exploratory thread: claim audit, live demos, and runnable code
  _{Benchmark · ★3 · samuelsacco · Py}

  Caveats: no licence

• jev-plays
  A System One model plays Craftax while an LLM sets the goals: five agents on the same map, from Jev on raw actions to an LLM controlling every step, compared in logged episodes.
  _{Benchmark · ★3 · mansicer · Py}

• origin-civilization
  AI life-and-civilization simulation: TypeSafe Jev makes every decision (typed, probabilistic, auditable); LLMs plan — OpenAI-compatible APIs, local models (Ollama, LM Studio), Claude Code, Codex.
  _{Benchmark · ★3 · jacquesgariepy · TS}

• typesafe-jev-calibrate-for-code-review
  About calibrating Jev for code reviews
  _{Benchmark · ★3 · selmar · Py}

  Caveats: no licence

• jev-agent-failure-benchmark
  Benchmarking Jev (Typesafe.ai) against a strong LLM on the Who&When Pro agent-failure-attribution benchmark (text subset).
  _{Benchmark · ★2 · tokentrim · Py}

• jev-play-ping-pong
  Jev plays browser table tennis in real time: structured telemetry, typed decisions, ordinary Chrome inputs, and auditable evidence.
  _{Benchmark · ★2 · icohen007 · JS}

• jev-routing-experiment
  Benchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArena
  _{Benchmark · ★2 · tokentrim · Py}

• zerosweep
  Autonomous System-One Triage Engine & Benchmark powered by TypeSafe AI (Jev). 75ms inference, $0 output tokens, and RLCD epistemic safety gates.
  _{Benchmark · ★2 · sysadarsh · TS}

  Caveats: no licence

• antigravity-mcp-semantic-search-with-typesafeai
  Fast semantic code search & diff sanity auditor for AI coding assistants (Antigravity, Cursor, Claude Code) powered by TypeSafe System One.
  _{Benchmark · ★1 · greenyamao · Py}

  Caveats: no licence

• can-jev-bayes
  Jev Bayes, No? Testing TypeSafe AI's Jev against Bayesian-optimal strategies, and testing if Jev can effectivly use Bayesian priors.
  _{Benchmark · ★1 · tomrichner · Py}

• decision-bench
  Open benchmark runtime for document-grounded decision models
  _{Benchmark · ★1 · hanno-labs · Py}

• dsh-jev-verify
  Jev (TypeSafe System One) decision tools + live verification benchmark for DeepSeek Harness: jev_decision (choice/score/noul) and jev_verify, honest by design.
  _{Benchmark · ★1 · xienda · JS}

• jev-bench
  Does the cited source actually say it? A 42-claim benchmark: Jev (TypeSafe System One) against GPT-5.4, Claude Sonnet 5 and Gemini 3.1 Pro.
  _{Benchmark · ★1 · thewaywithin · Py}

• jev-benchmark
  Reproducible benchmark for TypeSafe AI's Jev on agent tool-call risk classification: accuracy, latency, and whether the confidence score is worth routing on.
  _{Benchmark · ★1 · themsquared · Py}

• jev-decision-benchmarks
  JEV decision benchmark results on MetaTool, When2Call, and BFCL V4, with bilingual tables and reproducible reports.
  _{Benchmark · ★1 · baibizhe · Py}

  Caveats: no licence

• jev-eval
  Benchmark TypeSafe Jev against any OpenRouter model on your own labelled classification data: accuracy, calibration, latency, cost
  _{Benchmark · ★1 · 4esv · Py}

  Caveats: no licence

• jev-lab
  Hands-on research lab for TypeSafe's Jev (System One model): reproducible benchmarks of Noul/Choice/Score primitives, confidence gating, fan-out latency, agent control — plus a living audit of the Jev ecosystem.
  _{Benchmark · ★1 · llt22 · Py}

  Caveats: no licence

• jev-lab
  Experiments on TypeSafe Jev (System One decision model) via OpenRouter: repeatability, perturbation, and LLM baseline comparison
  _{Benchmark · ★1 · danielhirt · TS}

  Caveats: no licence

• jev-secret-detection
  Measures how well TypeSafe's RLCD-Jev model spots real secret credentials in file snippets
  _{Benchmark · ★1 · teyhouse · Py}

  Caveats: no licence

• jev-sim
  Jev-compatible /v1/systemone server reading typed decisions from LLM logits, benchmarked against TypeSafe's Jev on the same items via JevBench
  _{Benchmark · ★1 · dashbi1 · Py}

• jevsbistro
  3D restaurant service simulator for benchmarking low-latency decision models
  _{Benchmark · ★1 · andrewsilber · TS}

• padflow-jev-evals
  Typed-decision benchmark from PadFlow (land development SaaS): schemas, anonymized labeled rows, and a runner for confidence-calibrated models like TypeSafe Jev.
  _{Benchmark · ★1 · zsavage8 · Py}

• what-is-jev
  Independent, source-linked research on TypeSafe AI's Jev (System One), with 947 rubric-scored public repositories, recurring patterns, datasets, and bilingual documentation.
  _{Benchmark · ★1 · g0runmezadam · Py}

  Research about the ecosystem rather than a caller of the API, so it carries no call-site evidence.

• agent-handoff-gate
  An experimental protocol for evidence-aware agent handoffs, bounded worker continuation, and TypeSafe/Jev-assisted review, with reproducible evaluation.
  _{Benchmark · ★0 · zsoxi · Py}

• jev-acento
  ¿Jev entiende tu acento? Pre-registered audit of TypeSafe AI's Jev on Spanish — accuracy, calibration and token cost — plus a CLI to run the same comparison on your own labelled data.
  _{Benchmark · ★0 · marcosmartinez · Py}

  An independent, pre-registered audit of the model outside English — the gap docs/status.md lists as worth watching.

• jev-calibration-audit
  Independent API-only calibration audit of TypeSafe AI's Jev decision model
  _{Benchmark · ★0 · jujumilk3 · Py}

• jev-certify
  Finite-sample guarantees for Jev (TypeSafe's System One). Conformal risk control turns calibrated probabilities into certified routing thresholds; prediction-powered inference audits them. 2,412 decisions on CLINC150 for $0.23 — including the shift and prevalence cases where the guarantee break
  _{Benchmark · ★0 · nikkoxgonzales · Py}

• jev-cyrillic-audit
  Does TypeSafe's Jev keep its accuracy and calibration on Russian? Independent RU vs EN audit (ECE, reliability diagrams, paired bootstrap) on parallel human-labelled data.
  _{Benchmark · ★0 · ahtoooxa · Py}

  An independent calibration audit outside English — the gap docs/status.md lists as worth watching.

• jev-enterprise-decision-fabric
  Architecture for running many semantic decisions through one validated path, with a labelled 111-case benchmark comparing TypeSafe Jev against a Claude baseline, and a dashboard for inspecting any single decision. Experimental, not production.
  _{Benchmark · ★0 · ghubnab99 · C#}

• jev-eval
  TypeSafe's Jev vs gpt-5.4-mini and gpt-5.6-luna on four public classification sets: cases, per-item answers, scoring, charts
  _{Benchmark · ★0 · onlyoneaman · TS}

• jev-eval
  A third-party check of Jev against two LLMs under identical conditions: routing booking inquiries to a photo-shoot service for tourists in Japan, sixty synthetic messages in four languages.
  _{Benchmark · ★0 · shogo-nfrealmusic · TS}

  Caveats: no licence

• jev-fanout-bench
  Measured: asking TypeSafe Jev N questions in one call bills the state once. 2,976 real requests, raw data, exact billing check.
  _{Benchmark · ★0 · blowxian · Py}

• jev-lab
  TypeScript experiments, evaluations, and latency benchmarks for TypeSafe's Jev model
  _{Benchmark · ★0 · menny1337 · TS}

  Caveats: no licence

• jev-llm-router-benchmark
  Benchmark-driven Jev router and judge for cost-aware, reliable LLM coding workflows
  _{Benchmark · ★0 · erendikmenn · Py}

• jev-no-enem
  Reproducible benchmark evaluating TypeSafe AI's Jev (System One paradigm) on Brazil's ENEM 2025 standardized exam. Evaluates typed decision-making, domain-specific accuracy, and RLCD uncertainty calibration against open LLM baselines with an interactive GitHub Pages dashboard.
  _{Benchmark · ★0 · patryckalves · Py}

  Caveats: no licence

  An independent evaluation outside English, on a public exam with known answers.

• jev-orderby-bench
  Does ORDER BY over a Jev probability put rows in a defensible order? Independent ranking, calibration and invariant measurements of TypeSafe AI's Jev: passes six pre-registered gates on 360 labeled rows, fails four of six on graded product relevance.
  _{Benchmark · ★0 · yodablocks · Py}

• jev-playground
  Benchmarks Jev against other evaluation models in games with explicit states and legal actions: code owns the rules and transitions, each model picks the next action, and outcomes are measured.
  _{Benchmark · ★0 · hegargarcia · TS}

  Caveats: no licence

• jev-trace-classifier
  Application of TypeSafe Jev (noul judgment primitive) on the collusion.wiki corpus: agent vs human page authorship, head-to-head vs local Qwen3.8-Flash-Next
  _{Benchmark · ★0 · sypherin · Py}

• smoking-extraction-benchmark
  Synthetic smoking-history extraction benchmark comparing TypeSafe Jev and OpenAI structured outputs, with reproducible accuracy, cost, and latency results.
  _{Benchmark · ★0 · vclic · Py}

  Caveats: no licence

• An early-access test of TypeSafe's Jev: calibrated judgments for half a cent
  The best independent test found: 24 Norwegian documents on one pinned model version, opening with a case the model got wrong while correctly reporting low confidence.
  _{Benchmark · Lindfors}

  Methodology is stated cleanly and scoped honestly as a single-day snapshot. Leading with a failure case is what makes it a real calibration test rather than a testimonial.

• Testing TypeSafe Jev, Mistral and Gemini for local event validation
  The only three-way head-to-head found, with each model's prompt tuned separately and the scope limited to one task rather than a general ranking.
  _{Benchmark · Near Here}

  Self-limits correctly: a use-case study, not a model leaderboard. That restraint is rarer than the numbers.

The primary index. Each heading is a decision an agent has to make; the rows are examples of making it. Caveats appear as short tags — the full note for each row is in catalog.json and on the site.

Which tool or action the agent should call next.

• Cookbook: Function calling ⭐
  Maps natural-language trading requests onto ordinary typed functions by turning function names and closed-set arguments into confidence-aware questions.
  _{Official docs · Py · choice}

• Cookbook: Skill suggestion ⭐
  Picks at most one skill out of 182 for an agent turn: one request ranks every skill and asks whether the turn needs one at all, a second reads the top three.
  _{Official docs · Py · choice · noul}

• Demo: Smart home assistant ⭐
  Runnable demo code for a smart home assistant that evaluates user requests with typed decisions.
  _{Official docs · Py}

• ai-hedge-fund
  An AI Hedge Fund Team
  _{Integration · ★63,705 · virattt · Py}

• claude-code-templates: three Jev plugins
  Three independently installable Claude Code plugins — guardrails, model router and skill suggestion — each with its own hooks and tests.
  _{Plugin · ★31,582 · Py · TS · choice · score · noul}

• Composio TypeSafe provider
  Compiles a tool catalogue into questions and reconstructs tool calls from the answers, with typed errors for abstention and confirmation-required cases.
  _{Project · ★30,300 · Py · choice}

• FastMCP jev_search transform
  Two-stage MCP tool search: a wide Choice coarse-ranks the whole catalogue, then a shortlist gets full descriptions plus one Noul each to decide whether it does the job at all.
  _{Project · ★27,886 · Py · choice · noul}

• Cua driver: jev-use example
  Computer-use action selection in Python and TypeScript: Jev picks the next browser action from an immutable candidate set, with reobserve and abstain as reserved options.
  _{Project · ★26,164 · Py · TS · choice}

• jev-ultrafast
  A high-speed browser agent from Browser Use: Jev decides the operation and which element to act on, and a small LLM is called only when text must be typed.
  _{Project · ★19,261 · Browser Use · Py · choice}

  Caveats: vendor numbers

• json-render
  Vercel Labs' generative UI framework. In its Jev experiment the model does not write JSON token by token — it only picks components, props and layout.
  _{Project · ★18,204 · Vercel Labs · TS · choice}

10 of 230 shown · all 230 on one page → · filter on the site

_{↑ Pattern index}

Classify what the user wants and send the request down the right branch.

• Demo: Smart home assistant ⭐
  Runnable demo code for a smart home assistant that evaluates user requests with typed decisions.
  _{Official docs · Py}

• Pattern: Confidence-gated routing ⭐
  Treat confidence as a second axis: the answer tells you what, the confidence tells you whether to act on it.
  _{Official docs · Py}

• Pattern: Intent routing ⭐
  Classify an incoming request and route it to the cheapest adequate handler: deterministic code, a specialist LLM, or a person.
  _{Official docs · Py · choice}

• AutoGPT TypeSafe blocks
  Seven production blocks — choice, score, yes/no, ask-many, route, pick-best, filter — with a UTF-8 byte budget, verbatim wire capture and eleven test files.
  _{Project · ★187,515 · Py · choice · score · noul}

• Airflow LLMBranchOperator with Jev
  Turns downstream task ids into a choice option set, with a minimum-confidence gate that routes uncertain runs to a human.
  _{Integration · ★46,958 · Py · choice}

• Inbox Zero: seven email decisions
  Seven distinct email decisions, each with its own separately chosen threshold, falling back to the normal LLM on any error.
  _{Project · ★12,328 · TS · choice · noul}

• jev-chat-jarvis
  An Android reply co-pilot that judges intent, timing and risk from on-screen text, while separate models handle OCR and drafting.
  _{Project · ★5,414 · Java · choice · score · noul}

• Real Python: hello-jev
  A teaching example with a deliberate control group: the same station-enquiry task written in plain Python that only accepts Y/N, next to a Noul that reads intent.
  _{Tutorial · ★5,205 · Real Python · Py · noul}

• ai-cookbook: Jev track
  A graded course from a first call through each primitive, state shapes and criteria, to ticket triage and a multi-step workflow, mirroring all four official patterns.
  _{Tutorial · ★4,578 · Py · choice · score · noul}

• foreman
  A software-factory foreman that uses Jev to decide what an agent pipeline should do next.
  _{Project · ★538 · thruwire · Py}

10 of 35 shown · all 35 on one page → · filter on the site

_{↑ Pattern index}

Decide which tool calls and results still matter so stale context can be dropped.

• Hermes Agent: Jev compaction evaluation
  Ported the Jev compaction approach, measured it against their shipping summariser, and published the conclusion not to adopt it.
  _{Benchmark · ★248,479 · Py · noul}

• jcode: memory recall without embeddings
  Replaces the whole retrieval stack for memory recall — no embeddings, no BM25, no reranker — with one batched Noul per candidate memory.
  _{Project · ★20,069 · Rs · noul}

• fast-jev-compaction
  A Claude Code plugin that replaces the compaction summary with per-item decisions: stale tool calls are dropped or truncated, everything kept stays verbatim.
  _{Plugin · ★6,616 · tamaratran · TS · noul}

• hermes-jev-skills
  Nine agent skills plus a CLI covering model routing, memory filtering, turn retention, one-of-many skill selection and next-action choice.
  _{Plugin · ★718 · Py · choice · score · noul}

• compact-adviser
  "Work appears completed or recorded. Run /compact to save tokens."
  _{Project · ★183 · kunchenguid · TS}

• jev-pruner
  Trims long shell output before the model sees it, asking one Noul per chunk.
  _{Plugin · ★144 · tamaratran · TS · noul}

• Winnow
  Context garbage collection for Claude Code: when Read, Bash or Grep dump a wall of output, each chunk is judged for relevance to the current task.
  _{Plugin · ★79 · Py · noul}

• save-token-jev-clean
  Portable, Jev-guided context compaction for coding agents: instead of an LLM rewriting old context into a lossy summary, Jev decides which tool calls and results still matter, and user and assistant text is kept verbatim.
  _{Plugin · ★69 · iamunbounded · TS}

• yoshi
  Context-pruning proxy for Claude Code and Codex: Jev judges which history is still needed, measured not claimed. POC here now, heading soon into https://github.com/compozy/compozy
  _{Plugin · ★25 · compozy · TS}

• claude-jev
  Claude Code plugin: Jev for rule checks, verbatim compaction, and prompt routing
  _{Plugin · ★12 · 0x7067 · Py}

10 of 34 shown · all 34 on one page → · filter on the site

_{↑ Pattern index}

Decide whether an action is safe to run. Defence in depth, never a security boundary.

• Cookbook: Classifying RAG passages ⭐
  Scores each retrieved passage, then decides in code which reach the answering model — keeping contradictory ones flagged and dropping ones carrying prompt injection.
  _{Official docs · Py}

• Cookbook: Guardrails for LLMs ⭐
  Screens every message in and out of an LLM app in one request, naming hazards and scoring how much harm complying would do.
  _{Official docs · Py · noul · score}

• sub2api: Jev as a moderation endpoint
  Drops in as a moderation API by asking many parallel Noul questions in one request, one per hazard category, with an anti-injection prefix on every instruction.
  _{Project · ★42,557 · Go · noul}

• claude-code-templates: three Jev plugins
  Three independently installable Claude Code plugins — guardrails, model router and skill suggestion — each with its own hooks and tests.
  _{Plugin · ★31,582 · Py · TS · choice · score · noul}

• @langchain/typesafe
  The JavaScript counterpart of the LangChain integration, with the same classifier and middleware shapes.
  _{Integration · ★18,223 · TS · choice · score · noul}

• DeepChat: agent tool-permission review
  Reviews each tool call on three axes — risk level, whether the user authorised it, and an explicit prompt-injection pressure check.
  _{Project · ★6,341 · TS · choice · noul}

• agentgateway: CI-validated LLM guardrail
  Three Score questions on a shared severity scale, blocking the request when two or more cross the line, and failing closed.
  _{Project · ★5,017 · Rs · score}

• atomic
  The verifiable coding agent runtime. Define your coding agent's process in natural language with stages, checks, and approval gates instead of hoping it follows your instructions.
  _{Project · ★820 · bastani-inc · TS}

• Jev-cu
  A computer-use agent that asks which accessibility-tree element to act on, plus a separate noul for whether the action needs explicit user confirmation.
  _{Project · ★591 · JS · choice · noul}

• vexjoy-agent
  VexJoy AI Agent with Jev Intelligent Routing - /do routes plain-English requests to the right specialist agent and gates the work with reviews, tests, and a learning loop.
  _{Project · ★425 · notque · Py}

10 of 138 shown · all 138 on one page → · filter on the site

_{↑ Pattern index}

Check a model's output against a rubric before it reaches a user.

• Cookbook: Double-checking citations ⭐
  Catches wrong or invented citations against the source document with one Choice, using its confidence to flag borderline cases for review.
  _{Official docs · Py · choice}

• Cookbook: Guardrails for LLMs ⭐
  Screens every message in and out of an LLM app in one request, naming hazards and scoring how much harm complying would do.
  _{Official docs · Py · noul · score}

• latitude-llm
  Open-source observability for AI agents. Find where your agents fail, dispatch your coding agent to fix it, and verify the fix against real traces.
  _{Project · ★4,672 · latitude-dev · TS}

• reticle
  AI agents can generate code, but still struggle to understand what they build. Reticle brings Jev-style machine-native runtime perception to web & desktop applications.
  _{Project · ★830 · reticlehq · TS}

• atomic
  The verifiable coding agent runtime. Define your coding agent's process in natural language with stages, checks, and approval gates instead of hoping it follows your instructions.
  _{Project · ★820 · bastani-inc · TS}

• vexjoy-agent
  VexJoy AI Agent with Jev Intelligent Routing - /do routes plain-English requests to the right specialist agent and gates the work with reviews, tests, and a learning loop.
  _{Project · ★425 · notque · Py}

• jev-mcp
  A ready-made judgement toolbox for agents: fact verification, content screening, semantic ranking, classification and extraction as separate tools.
  _{Plugin · ★320 · JS · choice · score · noul}

• JevRev
  The decision layer beside an LLM: Jev filters plans, checks progress and keeps attention on work worth continuing, while the LLM supplies breadth and implementation.
  _{Project · ★303 · alex314618-create · TS}

• jev-review
  A local-first MCP plugin for continuous code-quality review by coding agents.
  _{Plugin · ★217 · niazmorshed2007 · TS}

• abide
  Make your coding agent abide by all your project rules
  _{Plugin · ★211 · coldteadotai · TS}

10 of 134 shown · all 134 on one page → · filter on the site

_{↑ Pattern index}

Decide whether a failed step is worth retrying.

• jevswiftsdk
  An independent, type-safe Swift SDK for TypeSafe Jev, with async/await, batching, retries, and SPM support.
  _{SDK · ★8 · nsstudent · Swift}

• jev-harness
  Zero-dependency System One decision harness: 5 semantic gates saving frontier AI agent tokens on trivial errors & doom loops. Python + TypeScript + Rust. MCP-compatible.
  _{Plugin · ★6 · ismaelsoilet · Py}

• jev-resilience
  Non-blocking Spring Boot Starter for Spring WebFlux that implements a Semantic Circuit Breaker to detect silent HTTP 200 failures using TypeSafe Jev.
  _{Plugin · ★2 · vicente-md · Java}

  Caveats: no licence

• Jev by Example
  Ten runnable JavaScript agent decisions, one file each: reconciling a new memory against a stored one, gating whether an HTTP 200 really satisfied the task, retry vs. reconcile after an uncertain write, scoring context against a budget, checking a handoff for dropped prohibitions.
  _{Project · ★1 · Really Artificial · JS · choice · score · noul}

  Caveats: one commit · AI-written

• jev-reasoning-navigator
  JEV Reasoning Navigator: Cognitive supervision, loop prevention, and anti-hallucination engine for autonomous LLM agents using TypeSafe AI
  _{Project · ★1 · andreuvm · Py}

  Caveats: no licence

• harnessjudge
  Judge agent steps — ok / retry / escalate / stop via TypeSafe Jev
  _{Project · ★0 · ndolinschi · TS}

  Caveats: no licence

All 6 shown · on its own page · filter on the site

_{↑ Pattern index}

Use calibrated confidence to decide what a person must see.

• Cookbook: Classification using confidence ⭐
  Classifies annual reports into 75 industry groups, then reads the answer's own confidence to decide whether to report that group or the broader division above it.
  _{Official docs · Py · choice}

• Cookbook: Double-checking citations ⭐
  Catches wrong or invented citations against the source document with one Choice, using its confidence to flag borderline cases for review.
  _{Official docs · Py · choice}

• Cookbook: Knowledge graph entity alignment ⭐
  Decides which of 450 candidate pairs from two product catalogues describe the same thing, with one Score whose three levels are the three available actions.
  _{Official docs · Py · score}

• Cookbook: Self-consistency with choices ⭐
  Adds an explicit "uncertain" outcome to moderation decisions and measures label agreement against the share of actions taken automatically.
  _{Official docs · Py · choice}

• Cookbook: Self-consistency with nouls ⭐
  Routes uncertain probabilities to human review while keeping the underlying noul values visible rather than collapsing them to a label.
  _{Official docs · Py · noul}

• Pattern: Confidence-gated routing ⭐
  Treat confidence as a second axis: the answer tells you what, the confidence tells you whether to act on it.
  _{Official docs · Py}

• Confidence ⭐
  How confidence is derived from the probability distribution, and why a threshold tuned on one question type does not transfer to another.
  _{Official docs}

• Airflow LLMBranchOperator with Jev
  Turns downstream task ids into a choice option set, with a minimum-confidence gate that routes uncertain runs to a human.
  _{Integration · ★46,958 · Py · choice}

• Composio TypeSafe provider
  Compiles a tool catalogue into questions and reconstructs tool calls from the answers, with typed errors for abstention and confirmation-required cases.
  _{Project · ★30,300 · Py · choice}

• Inbox Zero: seven email decisions
  Seven distinct email decisions, each with its own separately chosen threshold, falling back to the normal LLM on any error.
  _{Project · ★12,328 · TS · choice · noul}

10 of 67 shown · all 67 on one page → · filter on the site

_{↑ Pattern index}

Pick which downstream model or tier should handle a request.

• Cookbook: Structured data extraction cascade ⭐
  A two-stage mini-then-verify-then-reasoning cascade that reaches most of a big reasoning model's quality at a fraction of the cost.
  _{Official docs · Py}

• Pattern: Intent routing ⭐
  Classify an incoming request and route it to the cheapest adequate handler: deterministic code, a specialist LLM, or a person.
  _{Official docs · Py · choice}

• claude-code-templates: three Jev plugins
  Three independently installable Claude Code plugins — guardrails, model router and skill suggestion — each with its own hooks and tests.
  _{Plugin · ★31,582 · Py · TS · choice · score · noul}

• @langchain/typesafe
  The JavaScript counterpart of the LangChain integration, with the same classifier and middleware shapes.
  _{Integration · ★18,223 · TS · choice · score · noul}

• hermes-jev-skills
  Nine agent skills plus a CLI covering model routing, memory filtering, turn retention, one-of-many skill selection and next-action choice.
  _{Plugin · ★718 · Py · choice · score · noul}

• jev-review
  Pre-screens code review with Jev to surface high-risk changes for a more expensive model or a person, with a local dashboard.
  _{Project · ★582 · TS · choice · score · noul}

• jev-codex-router
  Judges how hard a coding turn is, then picks the model tier, reasoning depth and speed mode to match.
  _{Plugin · ★260 · JS · choice · score}

• Astra-Ares
  Adaptive reasoning effort for GPT-6 during Codex tasks, powered by Jev to reduce token usage.
  _{Plugin · ★250 · miuuyy · JS}

• jevrouter
  A router for models, tools and subagents.
  _{Project · ★189 · billionsbobby · TS}

• jev-eval-agent
  An agent that routes evaluation work through typed decisions.
  _{Project · ★105 · vinilana · TS}

  Caveats: no licence

10 of 43 shown · all 43 on one page → · filter on the site

_{↑ Pattern index}

Pack many questions — including speculative ones — into one request and let code pick what mattered.

• Cookbook: Parallel questions ⭐
  A 13-question regulatory briefing over one long article, showing that batching every question into one call is far cheaper and faster with no change in answers.
  _{Official docs · Py}

• Pattern: Speculative fan-out ⭐
  Pack many questions, including ones you may not need, into a single request and let your code decide afterwards what was relevant.
  _{Official docs · Py}

• Quickstart ⭐
  The canonical first call: one support ticket, one Choice, one Score and one Noul in a single request, in Python, JS and cURL.
  _{Official docs · Py · TS · sh · choice · score · noul}

• AutoGPT TypeSafe blocks
  Seven production blocks — choice, score, yes/no, ask-many, route, pick-best, filter — with a UTF-8 byte budget, verbatim wire capture and eleven test files.
  _{Project · ★187,515 · Py · choice · score · noul}

• sub2api: Jev as a moderation endpoint
  Drops in as a moderation API by asking many parallel Noul questions in one request, one per hazard category, with an anti-injection prefix on every instruction.
  _{Project · ★42,557 · Go · noul}

• jev-ultrafast
  A high-speed browser agent from Browser Use: Jev decides the operation and which element to act on, and a small LLM is called only when text must be typed.
  _{Project · ★19,261 · Browser Use · Py · choice}

  Caveats: vendor numbers

• ai-cookbook: Jev track
  A graded course from a first call through each primitive, state shapes and criteria, to ticket triage and a multi-step workflow, mirroring all four official patterns.
  _{Tutorial · ★4,578 · Py · choice · score · noul}

• jev-chat: a tool-calling chatbot with no LLM
  A chat bot that does tool calling with no language model anywhere: one request asks the request kind, the tool, and every tool's arguments at once.
  _{Project · ★94 · TS · choice · noul}

• jev-sift
  Classify first. Read selectively. A portable agent plugin and MCP tool for batch text classification.
  _{Plugin · ★46 · kbhuw · JS}

  Caveats: no licence

• pi-typesafe
  TypeSafe decisions for Pi: batched evaluation tool, terminal playground, and typed API for extension authors
  _{Plugin · ★45 · devmortimer · TS}

10 of 32 shown · all 32 on one page → · filter on the site

_{↑ Pattern index}

Score or re-rank candidates from a cheaper retrieval step.

• Cookbook: Classifying RAG passages ⭐
  Scores each retrieved passage, then decides in code which reach the answering model — keeping contradictory ones flagged and dropping ones carrying prompt injection.
  _{Official docs · Py}

• Cookbook: Line-by-line search ⭐
  Semantic search over a terms-of-service document: one request scores 218 line ids with a Choice, and a Noul checks whether the document answers at all.
  _{Official docs · Py · choice · noul}

• Cookbook: Re-ranking ⭐
  Re-ranks 30-passage BM25 shortlists for 40 legal queries with one question per query-candidate pair, reporting large top-1 and top-10 gains.
  _{Official docs · Py}

• AutoGPT TypeSafe blocks
  Seven production blocks — choice, score, yes/no, ask-many, route, pick-best, filter — with a UTF-8 byte budget, verbatim wire capture and eleven test files.
  _{Project · ★187,515 · Py · choice · score · noul}

• OpenViking: retrieval reranking
  One Noul per candidate document in a single batched request, with the yes-probability used directly as the relevance score.
  _{Project · ★38,571 · Py · noul}

• FastMCP jev_search transform
  Two-stage MCP tool search: a wide Choice coarse-ranks the whole catalogue, then a shortlist gets full descriptions plus one Noul each to decide whether it does the job at all.
  _{Project · ★27,886 · Py · choice · noul}

• jcode: memory recall without embeddings
  Replaces the whole retrieval stack for memory recall — no embeddings, no BM25, no reranker — with one batched Noul per candidate memory.
  _{Project · ★20,069 · Rs · noul}

• LanceDB TypeSafeReranker
  A vector-database reranker that asks one Noul per result and uses the yes-probability as an absolute relevance score, comparable across queries.
  _{Project · ★11,513 · Py · noul}

• no-mistakes: Jev review pre-brief, measured and retired
  One Score per candidate file to pre-brief code review — measured twice, then removed: more billed input for essentially no wall-clock gain, and offline replay showed the candidate list could not reach where review findings land.
  _{Benchmark · ★8,617 · Go · score}

• jev-chat-jarvis
  An Android reply co-pilot that judges intent, timing and risk from on-screen text, while separate models handle OCR and drafting.
  _{Project · ★5,414 · Java · choice · score · noul}

10 of 64 shown · all 64 on one page → · filter on the site

_{↑ Pattern index}

Pull typed fields out of messy text by choosing among candidates rather than generating them.

• Cookbook: Date extraction ⭐
  Extracts absolute and relative dates by asking for the parts a document names, then resolving and validating them in code with confidence-based review.
  _{Official docs · Py}

• Cookbook: Pre-parsed value extraction ⭐
  Regexes find candidate emails, phone numbers and amounts; the model selects the requested span so code can normalise a verbatim value.
  _{Official docs · Py · choice}

• Cookbook: Structure recovery ⭐
  Reconstructs Markdown from plain text that lost its formatting, in two requests: one restitches hard-wrapped lines, one classifies every block.
  _{Official docs · Py}

• Cookbook: Structured data extraction cascade ⭐
  A two-stage mini-then-verify-then-reasoning cascade that reaches most of a big reasoning model's quality at a fraction of the cost.
  _{Official docs · Py}

• smart-paste
  Fills form fields from pasted text: the form's heading, labels and your text go to TypeSafe, and it inserts the values it matches for you to review before submitting.
  _{Plugin · ★38 · nomanjack · JS}

• jev-reviewer
  Data extraction for systematic reviews, quoted from the papers. Ask a trial report and its supplements your extraction form or a RoB 2, ROBINS-I, QUADAS-2 or TIDieR template; Jev points at the lines, every answer is a verbatim quote with its page, you check it and export the table. Files stay i
  _{Project · ★33 · choxos · JS}

• jev-macos-loop
  Open-source macOS AI computer use and native GUI automation on Apple silicon. Jev + OmniParser CoreML + Apple Vision OCR. Bring your own OpenRouter, Vercel AI Gateway, or TypesafeAI token.
  _{Project · ★21 · jcpsimmons · JS}

• jevfill
  A Chrome extension that fills web forms from unstructured notes with Jev: paste your details once as plain text, with no structured profile, then fill forms on demand.
  _{Plugin · ★18 · imohitmayank · TS}

• jeveryword
  Text extraction with Jev: field extraction, PII detection and exact quotes, built on TypeSafe's Jev.
  _{Project · ★4 · jkrup · JS}

• jev-mcp-dispatcher
  Natural-language MCP tool dispatcher powered entirely by TypeSafe's Jev — no general-purpose LLM. Discovers a simple MCP server's tool signatures at runtime and uses Jev's typed primitives (Choice/Noul) to pick the right tool and extract its arguments straight out of the sentence.
  _{Plugin · ★3 · abhishekashokvkumar · Py}

  Caveats: no licence

10 of 16 shown · all 16 on one page → · filter on the site

_{↑ Pattern index}

Put an item into a taxonomy, including deep hierarchies walked with probabilities.

• Cookbook: Classification using confidence ⭐
  Classifies annual reports into 75 industry groups, then reads the answer's own confidence to decide whether to report that group or the broader division above it.
  _{Official docs · Py · choice}

• Cookbook: Hierarchical classification ⭐
  Walks deep patent, retail, biomedical and source-code taxonomies with a parallel beam search over Choice probabilities.
  _{Official docs · Py · choice}

• Cookbook: Knowledge graph entity alignment ⭐
  Decides which of 450 candidate pairs from two product catalogues describe the same thing, with one Score whose three levels are the three available actions.
  _{Official docs · Py · score}

• Cookbook: Structure recovery ⭐
  Reconstructs Markdown from plain text that lost its formatting, in two requests: one restitches hard-wrapped lines, one classifies every block.
  _{Official docs · Py}

• worldmonitor: news threat classification
  Two Choice questions over threat level and category, held in shadow mode after a blind evaluation found Jev merely tied the incumbent model.
  _{Benchmark · ★87,298 · TS · choice}

  Caveats: shadow mode

• json-render
  Vercel Labs' generative UI framework. In its Jev experiment the model does not write JSON token by token — it only picks components, props and layout.
  _{Project · ★18,204 · Vercel Labs · TS · choice}

• Inbox Zero: seven email decisions
  Seven distinct email decisions, each with its own separately chosen threshold, falling back to the normal LLM on any error.
  _{Project · ★12,328 · TS · choice · noul}

• tax-doc-classifier
  Tax document page classifier built on Jev decisions. 100% strict accuracy across 261 IRS forms, ~$0.001 per page.
  _{Project · ★419 · kyotofin · TS}

• classifier-dev
  Zero-shot text classification over plain HTTP — no API key, no account. One Cloudflare Worker, a CLI, and an MCP server. https://classifier.dev
  _{Plugin · ★417 · mrmps · TS}

• docjev
  A very fast document classifier/splitter using Jev
  _{Project · ★414 · jerryjliu · Py}

10 of 119 shown · all 119 on one page → · filter on the site

_{↑ Pattern index}

Turn free text into numeric features for a classical downstream model.

• Cookbook: Autoresearch feature discovery ⭐
  An autoresearch loop that proposes questions, turns free text into numeric features, and uses model error to improve a supervised gradient-boosting regressor.
  _{Official docs · Py}

• nimble
  Local typed decisions, contrastive data curation, and model evaluation.
  _{Project · ★1,699 · bespokelabsai · Py}

  Caveats: no licence

• jev-align
  Builds calibrated decision functions from human feedback.
  _{Project · ★284 · sutro-sh · Py}

• Prism
  Does not place orders. It judges market conditions such as toxic flow and mean reversion, and hands the assessment to the existing strategy.
  _{Project · ★97 · TS · choice · score}

• jev-curate
  Curates training data: JSONL and Parquet rows are judged on quality, relevance and risk before deciding what reaches downstream training.
  _{Project · ★45 · Rs · score · noul}

• tiershift
  Shift every LLM call to the cheapest model that can handle it. Routing decided by TypeSafe Jev in ~180 ms. No training data. Policy in plain YAML. TypeScript and Python.
  _{Project · ★3 · iamvatsalpatel · TS}

• jev-board-lab
  Interactive explorer and Jev question workspace for Jev Board datasets.
  _{Project · ★0 · webgrga · JS}

  Caveats: no licence

• jev-calibrated-narrative-coding
  Calibrated conversion of police crash narratives into probabilistic crash variables with a System One model. Pipeline, schema and aggregated results.
  _{Project · ★0 · pozapas · Py}

All 8 shown · on its own page · filter on the site

_{↑ Pattern index}

Classify and route incoming documents, invoices and forms.

• tax-doc-classifier
  Tax document page classifier built on Jev decisions. 100% strict accuracy across 261 IRS forms, ~$0.001 per page.
  _{Project · ★419 · kyotofin · TS}

• docjev
  A very fast document classifier/splitter using Jev
  _{Project · ★414 · jerryjliu · Py}

• formanator
  Submit Forma https://joinforma.com benefit claims from the command line and Model Context Protocol (MCP) clients, with support for AI-powered receipt analysis with an LLM or Jev
  _{Plugin · ★99 · timrogers · Rs}

• doc-router
  A Document OCR Router to help route pages based on content.
  _{Project · ★26 · misbahsy · Rs}

• jev-capability-atlas
  Independent, evidence-based map of when TypeSafe's Jev actually holds up vs. breaks down — real API-call receipts, not a leaderboard. 中文為主的雙語 repo。
  _{Benchmark · ★26 · zaious · Py}

• jevmory
  Coding-agent memory where every fact is a verbatim quote graded by TypeSafe Jev's calibrated confidence. Local-first, SQLite receipts, zero dependencies.
  _{Project · ★9 · romiluz13 · Py}

• pdf-race
  Docling → Jev vs Docling → Gemini 3.8 Flash vs Gemini reading the PDF: same documents, one clock, scored against arXiv's own metadata
  _{Benchmark · ★9 · goodrahstar · JS}

• jev-document-classification
  JEV Document Classification enables the rapid and cost-effective classification of text-based documents using AI, leveraging TypeSafe's "System One" model.
  _{Project · ★4 · charlyhno-eng · TS}

• jev-builder
  A browser form for building requests to TypeSafe's Jev: pick a template, fill in the blanks, copy the request. No JSON, no install, runs locally.
  _{Project · ★3 · collapseindex · JS}

• decision-first
  Agent skill that spots bounded-judgment steps, tries a typed decision model (TypeSafe's Jev) first, and documents every attempt
  _{Plugin · ★2 · harrymunro · Py}

10 of 20 shown · all 20 on one page → · filter on the site

_{↑ Pattern index}

Route support tickets and conversations by intent and urgency.

• Quickstart ⭐
  The canonical first call: one support ticket, one Choice, one Score and one Noul in a single request, in Python, JS and cURL.
  _{Official docs · Py · TS · sh · choice · score · noul}

• ai-cookbook: Jev track
  A graded course from a first call through each primitive, state shapes and criteria, to ticket triage and a multi-step workflow, mirroring all four official patterns.
  _{Tutorial · ★4,578 · Py · choice · score · noul}

• spring-ai-typesafe
  A community Spring AI starter bringing typed decisions to Java, with a builder API over the three question types.
  _{Integration · ★36 · Java · choice · score · noul}

• jev-triage
  Message triage CLI in Go, built on the Jev decision model from TypeSafe AI. Categorizes messages, scores urgency, and flags low-confidence ones for human review.
  _{Project · ★3 · boldbug1 · Go}

• Example: three primitives in one request
  A minimal first call asking a choice, a score and a noul together, annotated with the asymmetries that catch people out.
  _{Snippet · Py · choice · score · noul}

  Caveats: code untested

• Jev AI Use Cases
  Walks through use case after use case — agent routing, an in-agent decision layer, ticket triage — each with a concrete option set and a sample response.
  _{Tutorial · Mehul Gupta · Py · choice}

  Caveats: paywall

• Jev on AI/ML API
  Another gateway route, notable because its endpoint path and request envelope differ again from both the native API and Cloudflare's.
  _{Integration · Py · noul · choice · score}

• Jev on Cloudflare Workers AI
  Workers AI binding and REST samples asking a noul, a choice and a score in one call, with the full response including per-answer confidence.
  _{Integration · TS · sh · noul · choice · score}

All 8 shown · on its own page · filter on the site

_{↑ Pattern index}

Score quality, risk or relevance on an ordered scale.

• Cookbook: Self-consistency with choices ⭐
  Adds an explicit "uncertain" outcome to moderation decisions and measures label agreement against the share of actions taken automatically.
  _{Official docs · Py · choice}

• Pattern: Composite scoring ⭐
  Break one broad judgement into atomic scores and combine them with weights that live in your code, not in the prompt.
  _{Official docs · Py · score}

• AutoGPT TypeSafe blocks
  Seven production blocks — choice, score, yes/no, ask-many, route, pick-best, filter — with a UTF-8 byte budget, verbatim wire capture and eleven test files.
  _{Project · ★187,515 · Py · choice · score · noul}

• worldmonitor: news threat classification
  Two Choice questions over threat level and category, held in shadow mode after a blind evaluation found Jev merely tied the incumbent model.
  _{Benchmark · ★87,298 · TS · choice}

  Caveats: shadow mode

• gptcache
  Semantic cache for LLMs. Fully integrated with LangChain and llama_index.
  _{Project · ★8,201 · zilliztech · Py}

• jev-chat-jarvis
  An Android reply co-pilot that judges intent, timing and risk from on-screen text, while separate models handle OCR and drafting.
  _{Project · ★5,414 · Java · choice · score · noul}

• ai-cookbook: Jev track
  A graded course from a first call through each primitive, state shapes and criteria, to ticket triage and a multi-step workflow, mirroring all four official patterns.
  _{Tutorial · ★4,578 · Py · choice · score · noul}

• jev-review
  Pre-screens code review with Jev to surface high-risk changes for a more expensive model or a person, with a local dashboard.
  _{Project · ★582 · TS · choice · score · noul}

• pg-jev
  A real PostgreSQL extension exposing the primitives as SQL functions, so a semantic decision can appear in a WHERE clause over any row type.
  _{Project · ★324 · Py · sh · choice · score · noul}

• llm2jev
  Adapt local language models into Jev-compatible structured decision engines with Choice, Score, and Noul outputs powered by prefill-only binary inference.
  _{Project · ★286 · yinsongxu · Py}

10 of 164 shown · all 164 on one page → · filter on the site

_{↑ Pattern index}

Choose what to surface next, fast enough for a live conversation.

• Jevflix
  Jev picks, you watch. A hybrid movie recommender: fast semantic + keyword search narrows 4,800 films to a shortlist, then TypeSafe Jev reads your constraints and picks the one film that fits - with a confidence score that decides whether to answer instantly or ask a follow-up.
  _{Project · ★0 · arielbubis · Py}

All 1 shown · on its own page · filter on the site

_{↑ Pattern index}

Surveys the model or the space rather than one pattern.

• Official agent skill for Claude Code ⭐
  Installs a TypeSafe skill into Claude Code so an agent can write correct Jev calls without you pasting the API shape each time.
  _{Official docs · ★2,036 · sh}

• typesafe-ai/skills ⭐
  The official agent-skills repository behind the Claude Code plugin, holding the SKILL.md that teaches an agent the System One API.
  _{Plugin · ★2,036 · sh}

• system-one-adapter-python ⭐
  A drop-in TypeSafeClient replacement backed by ordinary LLM APIs, so you can run Jev-shaped code without Jev access.
  _{SDK · ★285 · Py}

• @typesafe-ai/sdk (TypeScript / JavaScript) ⭐
  The official TypeScript client. Ships ESM, CJS and type declarations, with lowercase choice()/score()/noul() helper factories.
  _{SDK · ★232 · TS · JS · choice · score · noul}

• typesafe-sdk (Python) ⭐
  The official Python client. Sync and async clients, retry policy with retry-after support, and Choice/Score/Noul helper classes.
  _{SDK · ★219 · Py · choice · score · noul}

• API reference ⭐
  The one endpoint, POST /v1/systemone, with the exact request and answer shapes for all three question types.
  _{Official docs · sh · Py · TS}

• Models, pricing and limits ⭐
  The authoritative sheet: jev-1.13.0, $0.042 per Mtok input with output free, 64k context, 32k for state plus the longest question, text input only.
  _{Official docs · sh · Py · TS}

• Primitives: Choice, Score, Noul ⭐
  What each primitive is for and how to write criteria, including the 255-option cap on Choice and the 2-10 level range on Score.
  _{Official docs · Py · TS · choice · score · noul}

• Introducing System One models and Jev ⭐
  The launch post: what a System One model is, why decisions were split from generation, and the vendor's latency and cost claims.
  _{Article · Diogo Almeida}

  Caveats: vendor numbers

• Jev 1.13 known limitations ⭐
  The vendor's own list of where the model fails: literal reading, arithmetic and counting, date comparison, indirection, large noisy states, adversarial content.
  _{Official docs}

10 of 451 shown · all 451 on one page → · filter on the site

_{↑ Pattern index}

The same rows grouped by what you will find when you open the link.

The parts that are not the catalog.

• Link checks — 1204 rows carry an HTTP 2xx response and a checked date; 3 carry no dated success record. Dates vary by row and a past success does not guarantee availability today. Stars and licences are repository metadata snapshots.

• Source and code review — evidence.path cites the file read, evidence.read_on records the reported review date, and evidence_none explains missing file evidence. Reading a call site is separate from running it. Summaries include source descriptions and machine translations; see the method and its limits.

• Call-site text checks — 1121 rows record a file and matching strings in evidence. The weekly claims job checks that those strings remain on the default branch and reports missing text or files. This count measures recorded evidence, not latest CI passes. A text match does not prove that a call executes, the API is compatible, or the result is correct.

• Runtime and performance not independently tested here — treat every catalogue entry as untested by this repository, including entries without code-untested. Linked benchmarks describe their authors' measurements; this catalogue has not reproduced them. Repository build checks and package smoke tests do not exercise those integrations or the live Jev API, and inclusion is not a security review.

Links that stopped resolving, kept so a dead reference stays searchable instead of vanishing.

One entry per example, validated against a JSON Schema on every push.

Corrections take priority over additions — a wrong row costs more than a missing one. See CONTRIBUTING.md; the bar is could a reader act on this row without opening the link?

Code in scripts/, site/ and examples/ is MIT. Catalog metadata is CC0-1.0, with a per-row license field. Linked works keep their own licences — repo_license records what each declares.

Maintenance checks: · Scheduled link checks · Call-site text checks

Jev Decision Atlas · ↑ Back to top · 中文