Launches on September 22, 2026

Chrome extension that highlights Reddit posts you can answer, scored by TypeSafe Jev against your own expertise cards

Experimental local-first macOS voice automation using whisper.cpp, TypeSafe Jev routing, typed actions, and optional Ollama generation.

Automated resolution infrastructure API for long-tail prediction markets (Cloudflare Workers + Supabase + TypeSafe Jev)

Dockerized REST service, jev-compatible, that serves the Laya model (convaiinnovations/laya) for use from non-Python apps.

See your resume the way a screener sees it. Built on TypeSafe's Jev — typed judgments composed in code, not an agent loop.

Inspect failures, retries, and skipped jobs behind a green GitHub Actions run. Local evidence reports with optional Jev classification.

Open source Chrome extension for semantic RFx requirement checking, powered by TypeSafe Jev.

GitHub Action that auto-triages new rolldown issues with TypeSafe's Jev model: priority, reproduction quality, one comment.

Using System One modal Jev built by TypeSafe to check for scam related messages

Android app that warns you about scam messages (SMS, WhatsApp, Telegram, email) before you act on them, powered by TypeSafe's Jev model.

Search your Mac screen history. Native SwiftUI app with automatic Jev search, your own API key, and local fallback.

Selectia - A family of System One-style models fine-tuned from LFM 2.5, designed for one-pass typed decisions with calibrated probabilities. Decision API (Choice, Noul, Score) powered by open LLMs.

General-purpose decision models: choices, scores, and probabilities in one forward pass.

Describe a feature, get a verdict on what should actually power it: plain code, Jev, an LLM, Jev + LLM, or classical ML. An executable reading of TypeSafe's own jaggedness list.

Local MCP server giving coding agents Jev-backed filters: jev_grep, jev_rank, jev_search

Sieve: a Chrome extension that scores which LinkedIn and Reddit posts are worth your time and suggests reply angles, built on TypeSafe's Jev. Never posts for you.

Testing out Jev using the TypeSafe Python SDK

Evidence-backed AI answer-readiness auditing for web pages. Tells you what to fix, with a verbatim quote for every finding — never a score. Built on TypeSafe's Jev System One decision model, so no generative AI is used and no evidence can be invented. MIT.

JEV-inspired typed runtime decisions over long-horizon PSG: reusable overnight encoding, sparse retrieval, and Choice/Noul/Score probabilities.

Real-time AI slop detector for your feed, built by AdaL with TypeSafe Jev

Hides AI slop on X and LinkedIn using TypeSafe's Jev model. Typed judgments, code decides.

Jev（TypeSafe AI）で水平思考クイズの質問を判定するアプリ

AI-powered SQL Server plan analysis, query optimization, and Jev-ranked action plans.

Natural-language search over anyone's GitHub stars, re-ranked by TypeSafe Jev

Judgment layer for Stratify: asks TypeSafe's Jev model what static analysis could not prove, and moves each finding along the confidence ladder the engine already uses. Rust.

Benchmark for typed System One decision models (Jev, Laya, and whatever comes next): calibration against a noise floor, framing sensitivity, selective prediction, ordinal fidelity, interference, robustness. pip install sys1bench.

A Proxy that enables zero-shot system one typed-decisions endpoint for decoder only models served with vLLM or Sglang or Openrouter (Experimental)

Experiments with Jev, TypeSafe AI's first System One model

Real-time Telegram conversation analysis with Jev (TypeSafe System One). You write the draft, Jev judges it.

A semantic interrupt controller: calibrated semantic channels sampled over a live conversation at 5Hz, firing only on a hysteresis-gated edge. The model decides when; your code decides what. TypeSafe Jev + Deepgram Flux, zero dependencies.

Jev-powered model router - measured 77.8% token-cost savings on real workloads. Local-first, cloud when it earns its cost. No accounts, no keys.

Read-only Solana pump.fun screening desk with Terminal panels, GMGN charts, and Jev lore review. No trade execution.

Trier sa boîte mail avec Jev (TypeSafe) : 4 templates prêts à l'emploi - script Python, workflow n8n, scénario Make, skill Claude Code.

Machine-made, human-judged. Jev decides the design, an LLM writes the copy, an image model illustrates it — and code ships the page in about a second for a tenth of a cent. Typed decisions with calibrated confidence, not vibes.

Bench of typed decision models: Jev vs OpenJev vs Laya, small local LLMs and cheap hosted LLMs on the same zero-shot classification tasks, in English and French.

Unofficial typed wrapper around the TypeSafe AI SDK: ask typed questions, get unwrapped answers.

A simple SDK for integrating Typesafe AI's Jev into your Go projects

Unofficial, dependency-free TypeScript client for the TypeSafe System One API, emitted as a single readable ESM file.

typesafe-demo

TypeSafe AI Jev API Wrapper SDK

MCP for typesafe ai jev

Haskell SDK for TypeSafe AI

Notebooks built while learning TypeSafe AI's System One model Jev

Domain-agnostic neurosymbolic ODAV (observe/decide/act/verify) core

UiPath Maestro Flow x TypeSafe Jev - a card dispute triage bench comparing a calibrated typed-decision model against an inline AI agent on identical work, with latency and cost measured at runtime.

Minecraft (Paper 1.21) chat moderation plugin powered by JEV from TypeSafe

Interactive text signal observatory powered by TypeSafe Jev

An MCP server that gives AI coding agents an external "done" criterion — judged by TypeSafe AI's Jev decision model. Agents submit work before declaring completion and get specific, evidence-backed fix lists instead of a vague pass/fail.

Moderating live-stream chat in real time (ES/EN)

Assistant vocal français pour macOS : Whisper local + décisions Jev, actions sûres, interface Liquid Glass.