Typed decisions with TypeSafe's Jev, the first System One model
A website, two games, a benchmark and an agent skill for Laya, the open-source decision model. Runs on your machine.
Open-source, local-first social media research agent for Instagram, TikTok, and LinkedIn. Jev routes read-only steps; socai CLI captures cited browser evidence.
A foobar2000 theme: your albums as rendered specimen cases on shelves, a Möbius ring that pulses with the music, synced lyrics, 10 case skins × 5 colour schemes, light and dark.
Pluggable controls for reliable AI agents: validated outputs, bounded loops and frozen collective decisions.
Browser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.
Loop engineering for software development. Turn any repository into a self-organizing AI engineering team — 8 specialists, contract-driven, evidence-gated, parallel feature voyages.
Plain-English end-to-end tests for web apps. A small decision model picks each click; code decides pass or fail; a passing run exports a Playwright spec.
An open-source, efficient System I decision model for autonomous driving.
Agents can work. They cannot declare success. A governed coordination protocol for multi-agent AI work: eight contracts (task, resource, ownership, handoff, verification, evidence, approval, recovery) enforced in code, plus a reproducible conformance benchmark. Apache-2.0 code, CC BY 4.0 specification, zero dependencies, Node 18+.
🧹 Fun project: a Chrome extension that asks a tiny AI decision model (TypeSafe Jev) "is this DOM element an ad?" and pops it off the page. BYOK, no backend, not a real ad blocker.
DSH Community Consensus & Implementation Specifications / DSH 社区共识与实施规范
JEV-inspired parallel decisions for CUDA LLMs. One context, many decisions. vLLM API, game-agent examples, and reproducible benchmarks.
Train a 17M-parameter text classifier in a minute, with labels from a big AI model, and know when it beats an API. Recipe, tutorial, results on 10 tasks, phone trainer.
A source-reviewed gallery of JEV-related projects with 50+ GitHub stars — integrations, tools, open models, experiments, and ecosystem resources. Live gallery: beatapi.io/awesome-jev
Generate replayable Excel test cases from live browser pages, or run existing Excel cases, using Playwright and local Laya or a compatible decision API.
A calm family dashboard for a cheap Android photo frame: calendar, weather, Todoist, recipes, music and agent messages on a Cloudflare Worker. Inspired by Wes Bos's photo-frame-dashboard.
A high-performance, full-stack platform built specifically for the Kenyan software engineering community to discover, analyze, and countercheck software engineering roles across the worldwide work pool and local ecosystem.
Free multi-AI research assistant: your agent consults 6 top web AIs in parallel, cross-validates their answers, and generates an executable decision document. No API keys, no backend.
为 DeepSeek Harness 打造:选中 AI 回答或思考内容,随时追问、学习、检查与纠偏。
Circadia is an open OKLCH color specification for editors and terminals based on circadian science. It offers glare-free Warm Parchment for day and halation-free Warm Ember for night. These themes maintain strict WCAG 2.1 AAA legibility for continuous, low-strain focus during long writing and coding sessions.
briefme / brief-me: Turn complex AI analysis into a consultant-style, visual, interactive decision briefing.
Practical, tested recipes for TypeSafe's Jev decision model on OpenRouter: support triage, database indexing, file organizing, tagging, taxonomies, dedupe, PII detection, extraction, search re-ranking and a browser agent.
Dashboard local de tu patrimonio neto: fondos, ETF, acciones, cripto, pensiones y más. Precios automáticos, TIR, comparación con indexados e importación desde MyInvestor, Excel o con IA. Tus datos no salen de tu ordenador.
15 open-source outbound recipes on TypeSafe Jev. Score your LinkedIn network or any lead list against your ICP, catch job changes, triage replies. Local, zero dependencies. 16,711 connections scored for $0.73.
Laya (open source, local) vs TypeSafe Jev (API): two AI models race in Snake and fight in a Mortal-Kombat-style arena. Every move is a real model decision.
Special relativity simulator, with slow light speed.
An n8n workflow that finds local businesses to sell to, with a specific provable reason to call each one. Powered by AnyAPI: one key, USD pay-per-request.
麦麦全场景决策官是一个基于麦当劳中国 MCP 的点餐决策 Skill。它能整理单人或多人需求,比较门店、餐食、活动与优惠,核对逐人预算、总预算、忌口和活动场次,并给出分摊及代付方案。用户修改人数、预算或偏好后,它会保留仍有效的条件,重新核价并展示新旧方案差异。对无法核实的价格、权益和预约条件,它会明确标注未知;领券、兑换、预约和下单需另行确认。
Serve the official Ling-3.0-flash INT4 on one NVIDIA DGX Spark, 38.7 tok/s, faster than the community GGUF. Recipe, watchdog, benchmarks.