Jev ecosystem projects — page 152

Classify the Omarchy issue backlog with Jev, TypeSafe's System One model. Ten typed questions per issue in one request, for a hundredth of a cent each.

Native omp (oh-my-pi) extension: TypeSafe Jev judgment tools — token efficiency, confidence routing, citation verification

TypeSafe AI (Jev System One) plugin for OpenClaw - sub-100ms group triage, tool safety guardrails, compaction curation, and model routing

Optional typed TypeSafe AI Jev decisions for OpenClaw, with SecretRef credentials and strict API validation.

Open-source Jev-compatible MCP server for typed decisions (Choice/Score/Noul): local GLiNER backend with a model router, drop-in swappable with the TypeSafe System One API.

An agent in the browser that pays for thinking once: author a deterministic spec (Jev first, LLM on escalation), replay it for 0 tokens.

Agency-ready build prompt for a Personal Decision Inbox powered by TypeSafe AI (Jev). Full end-to-end spec: architecture, UI, integrations, deployment.

Shell command safety gate for the Pi coding agent, backed by TypeSafe's Jev judgment model

Experimental Pi extension using TypeSafe Jev for auditable tool-history compaction

TypeSafe Jev (System One judgments) for Pi: zero-dependency jev_ask tool with question linting, model discovery, and budget caps

High-concurrency Elixir actor runtime for large-scale semantic graphs, search swarms, and batched model evaluation.

Natural-language PostgreSQL predicates with TypeSafe Jev. A simple search playground with randomly generated databases.

Ask the Potpie docs a question, get the exact paragraph that answers it. Extractive search powered by Jev (TypeSafe System One) — no LLM, no embeddings, nothing generated.

Vite demo: TypeSafe Jev checks whether text contains profanity

An MIT-licensed public-radio assignment desk using NPR CDS MCP and TypeSafe Jev for editorial discovery, review, and audio workflows.

Black-box QA of saucedemo.com, judged three ways: TypeSafe Jev, GPT-5.6 with screenshots, and a confidence-gated hybrid.

X 时间线清洁工 · FeedSieve(MIT) 衍生 · 带实测标定的 AI 判定层：误杀 0.7%，还能抓词库认不出的 47%

Built a new SALT RSS pipeline that can do tellurics and such.

Synthetic smoking-history extraction benchmark comparing TypeSafe Jev and OpenAI structured outputs, with reproducible accuracy, cost, and latency results.

Jev on Indonesia's SNBT 2025 university entrance test: 159 questions, seven subtests, audited answer keys.