What people are shipping with Jev.

Most teams still use frontier models like a smarter intern. @typesafeai Jev is closer to a function. You don’t ask it to “think about the ticket.”
You ask: urgent? billing? needs h

Corent’s brain uses @typesafeai Jev now. Not to generate the output, to make the decision before the output gets generated. When a request comes into Corent, you don’t need to te

@unicodeveloper @typesafeai jev 🤝 us

@flarrs315 BOZO SCORE 91.7 / 100 종결 BOZO · 수익률 찡찡이 팔이 ████████░░ 78 정신승리 ██████████ 98 자아 █████████░ 92 공허 █████████░ 89 형식 █████████░ 91 "걍 찡찡거리는 병신" @typesafea

@lancehasson @typesafeai Yes! I actually started building it so that it exposes all in game actions via MCP / API so anyone can connect their agents to it. Then Jev came out and

@DSJayatillake @typesafeai I think Jev will be useful for sprinkling around the use cases and harnesses. not a full replacement at all. like, why can't Jev be part of a precommit

e2e + jev from @typesafeai ⚡ I'm building an open-source framework for running e2e tests with agents. supports web, mobile (and more!) available soon: https://t.co/G3XB2L8tiC htt

@ShenSeanChen @bot @Muse @SpaceXAI @AIatMeta @alexandr_wang Have you tried Jev by @typesafeai yet.

i don’t think i’ve had my feed ever stuck on one topic as much as the past 3 days with jev and @typesafeai . incredible motion

@altryne @typesafeai Is Jev GA already?

@gb3nga @typesafeai @ValyuOfficial To this particular app or Jev?

@studio_yebisu could you check out: https://t.co/VO0sYmPdzl its a jev driven coding agent ive been working on. lmk what you think!

Jev decides. Production needs the boring layer around it. OSS’d jev-harness, not affiliated with TypeSafe: policy → confidence gate → shadow → recipes → action evals Same 24-row

判断専用モデル Jev（TypeSafe）と汎用 LLM を比べるチュートリアル。 Jev の3つの問いの型、確信度での分岐、時系列の扱いを動くコードで。 ollama gemma4:12b + JSON schema でも形式は守れることに気づきました。 https://t.co/9KKAn6231H

哪位大佬可以试试AI打红警，优化下策略能打赢。https://t.co/D07gZgATpL 哪位大佬有Jev也试试 https://t.co/ZB7k5vJ9dA

Complementary layers. Browser Use owns the computer-use loop. We own the production decide layer when candidates are already structured: policy, confidence, shadow, evals. Claude C

@mattdesl @rmdnmd https://t.co/Dl2v9VzPV7 couldn't help myself. it's here. https://t.co/yWNDzxv0ag @mattdesl happy to host it under your name/domain and use my jev token

@mertcobanov Multiprovider cli üzerinden aboneliklede çalışan bir aracımız var kendimizde kullanıyorduk jev de entegre ettik deneyebilirsiniz yeni paketide npmden yayınlanacak bira

I built pi-heed for Pi. “Don’t modify files.” “Don’t touch tests.” “Never call production.” pi-heed makes sure those constraints still matter 20 tool calls later. https://t.co/M

@carloAI not trying to murder Pinterest 😅 you can run it here: repo: https://t.co/1QSgiCLO3f preview (no Jev): https://t.co/TADALFsDVQ hosted preview is gallery + search only