Jev

production GEO platform whose NOTRA_JEV_CLASSIFIERS flag routes brand-visibility classifiers off an LLM and onto Jev Boolean decisions at a 0.5 threshold, targeting 300 ms p50.

routes Claude Code tasks to the cheapest capable model by asking Jev to choose among candidates.

LiteLLM-based router where a Jev decision picks which model serves each request.

adds automatic per-request model routing to the Pi coding agent through Jev decisions on Vercel AI Gateway.

local proxy that picks the Claude model and reasoning effort per message with a Jev decision while leaving the cached main chat untouched.

routes agent skill selection through typed, confidence-aware Jev decisions so weak matches are declined instead of guessed.

screens a folder of CVs with Jev typed judgments against an editable policy, re-scoring candidates for free when the policy changes.

async LangGraph workflow gets a typed Jev Choice (invoice or general) and routes each inbound email to the matching handler.

WXT extension where Jev decides per page element whether it is clutter, removing it under reusable template rules.

staged code-review workflow and local dashboard where Jev gates each review stage before a change advances.

adds a measured tool-call gate to the Pi coding agent so risky calls are checked by Jev before execution.

wires Jev into its eval testkit as a verification judge so agent-produced work is gated by typed verdicts rather than a text model.

prompt-injection and dangerous-action guard for Claude Code, Codex, Pi, and ACP agents, with Jev deciding what to block.

scores every file of a pull request on 31 boolean Clean Code smells plus function size and nesting, then hands the verdicts to a writing model for the review prose.

checks whether each cited paper actually supports the sentence citing it, with Claude locating the quote, Jev scoring the support, and a human making the final call.

finds link paths between English Wikipedia articles by having Jev rank each page's outgoing links while Python controls the search.

browser-use's ultrafast agent where Jev decides each next action and element to click, calling a language model only when text must be typed.

exposes System One judgments as five Pi tools so a model makes narrow semantic judgments while code and users keep control of thresholds, weights, and actions.

agent skill that sends closed coding-agent judgments to Jev so verdicts stay typed, cheap, and comparable across runs.

Stop hook that keeps an agent from finishing too early by judging plain-language completion rules with Jev.

SO-101 arm workbench where a Jev decision runner picks bounded joint steps from typed candidate actions under a spend budget.

multiplayer triage dashboard where Jev answers a fixed set of typed questions per issue - kind, severity, urgency, duplicate, and next step - and every human correction is kept and shown back to the model on later runs.

benchmarks Jev against Luna, Haiku, and Gemini at choosing validated legal moves in explicit-state games, scoring decision quality and consistency across a sequence of moves.

reproducible eval harness plus field note for Jev Ultrafast research-browser tasks, with QC'd cases, a suite runner, and a report generator.

reproduces the System One shape with a Qwen3.5-2B fine-tune that emits typed decisions with calibrated probabilities in one pass.

independent local preview that answers bilingual probability questions from context, questions, and candidate answers, inspired by TypeSafe Jev.

RLCD-trained Qwen2.5-1B demo exploring parallel constrained decoding as an alternative to Jev.

Vercel's eve engine ships Jev as the default evaluation model (typesafe-ai/jev) in its experimental evaluate path.

Vercel Labs CLI that can run Jev as the evaluation model for its evaluate command.

proof-of-concept MCP server that puts Jev claim verification, content screening, and candidate ranking behind standard MCP tools.

MCP server exposing Jev classify, score, check, match, and screen as tools for any agent, with confidence on every answer.

ZIO client for TypeSafe AI with a typed DSL over Jev decisions.

unofficial Laravel integration for Jev with typed responses, async requests, scoped dependency injection, and testing fakes.

small type-safe client for asking Jev questions about a dataset.

async client for Jev published on PyPI.

Neon Function proxy for the Neon AI Gateway that routes decisions with Jev.

TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state, choosing each action from emulator-derived features.

camera-only autonomous drone in MuJoCo that puts a Jev judgment model in the control loop at 2.5 Hz.

drives original StarCraft shareware through keyboard and mouse with Jev action probabilities recorded per decision.

benchmark that asks Jev to predict federal motion-to-dismiss rulings from the judge's written record and scores the calibrated probabilities with claim-defendant micro-Brier metrics.

Discord bot that scores incoming messages for phishing, spam, and social engineering with Jev and drives a four-stage escalation ladder, injecting pardoned messages back into context as verified-safe precedent.

zero-shot spam classification with Jev Boolean questions, benchmarked against TF-IDF baselines.

Return typed Jev judgments and probability cards inside Agent Zero.

Trace TypeSafe System One requests and typed results in Braintrust.

Classify candidate DOM elements as advertisements in a browser-extension experiment.

Suggest existing shell-history commands by semantic fit to what the user has typed.

Label the tone of social posts and drafts and optionally hide matching posts.

Compare native Jev decisions with alternative models on shared synthetic application contracts.

Compile tool schemas into Jev questions and turn typed decisions into executable or partial tool selections.

Expose categorical Jev decisions as PostgreSQL scalar and batched SQL functions.

Turn semantic judgments about text into deterministic compiler-style diagnostics.

Triage newly opened GitHub issues and identify missing bug-report details.

Route code-review attention and suggested reviewers from a unified diff.

Deliver asynchronous Jev decisions to Elixir GenServer callbacks.

Apply typed judgments to CLI filtering, triage and tool-call risk checks.

Route coding tasks between two harnesses using task complexity and uncertainty.

Provide a typed tiebreaker CLI for Hermes agent workflows.

Reduce irrelevant tool output and select task-relevant memory for coding agents.

Grade readable page sections for writing, clarity and on-page SEO.

Detect error-like content in otherwise successful reactive HTTP responses.

Rerank vector, full-text and hybrid search results by semantic relevance.

Rank items using user-selected relevance factors and weights.

Explore typed move selection with two Jev players and a deterministic Gomoku board.

Choose high-level goals and game-button actions from emulator state.

Blur private page values during screen sharing using deterministic detectors and semantic classification.

Judge coding-agent tool calls after deterministic permission rules.

Route web searches to sources and time windows, then rank returned links.

Rank support tickets using a common urgency rubric.

Expose typed judgments about home state as Home Assistant sensors and automation inputs.

Check local Git changes against Markdown coding rules and show localized editor findings.

Add semantic rules to structurally validated Zod data.

Check coding-agent actions and edits against risk signals and project rules.

Use Jev-selected actions to drive a Hyprland desktop from OCR and window state.

Expose typed Jev evaluations as a stdio MCP tool for agents.

Route typed DSPy signature outputs to Jev while leaving free-form fields to a generative model.

Expose batched typed judgments to Pi agents, terminal users and extension authors.

Search for relevant code files with weighted walkers over the filesystem.

Wrap typed Jev decisions in an agent loop with tools, state, confidence policy and telemetry.

Select Minecraft skills, targets and exploration directions from structured game state.

Drive the ego lite browser using an indexed action space and typed Jev decisions.

Choose browser actions from a DOM snapshot and assess high-impact clicks.

Assess coding-worker progress and choose when deterministic supervisor policy should intervene.

Proxy native TypeSafe requests with shared authentication, logging and usage accounting.

Inspect source-code properties and changes with advisory Jev assessments.

Use semantic predicates, choices and scores over PostgreSQL rows.

Demonstrate a model-assisted SAFE/DANGER check before SQL migrations.

Turn partial voice transcripts into bounded browser actions.

Score transcript sentences and render probability meters into a video.

Explore resource activation and incident-log relevance in two small demos.

Map structured output fields and tool routing into typed Jev questions inside Pydantic AI.

Submit typed evaluations from shell scripts without a language SDK.

Filter, classify and rank SQL rows with natural-language predicates.

Select legal Snake moves from a deterministic game-state analysis.

Choose browser actions and decide when a task is complete or stuck.

Navigate a Neo4j graph toward a goal using a typed decision at each hop.

Display a typed risk assessment for a public GitHub pull request.

Resolve ambiguous Japanese addresses against a local address master.

Expose TypeSafe decisions through the AI SDK evaluation-model interface.