Jev AI examples, page 17

@joaobnobre shows TypeSafe Jev playing Minecraft Skyblock: one agent reads game state and chooses its next move through typed decisions.

Got @typesafeai’s Jev playing Minecraft Skyblock. One agent, reading the game state and choosing its next move through typed decisions. https://t.co/RXmybNhVtB

@dimentary tested Jev as a real-time robotics policy in MuJoCo. It struggled at first, so each update became two calls: what to do next, then how to move the arm and gripper. Jev gets simplified geometry and contacts as text, not images.

tested Jev as a real-time robotics policy in MuJoCo it struggled at first, so i split each update into two calls: decide what to do next, then decide how to mov…

@theoleecj points people still on the Jev waitlist to an alternative they call close enough, with a browser demo. The post does not name the model behind that demo.

On the Jev waitlist? You can use not Jev (but close enough) right here: https://t.co/S0EQnHkSGL https://t.co/qLtRs0fHyD browser demo https://t.co/BUFN88boT4

• github.com
• openjev.com

@vikktorrrre says Jev breaks a large prompt into smaller decisions you combine in your own code. The examples are: is this urgent, is this billing, and does a human need this.

we spent 3 years learning how to write prompts Jev from @typesafeai breaks big prompts into smaller decisions you can build around instead of one giant prompt h…

Quoted @CompleteSkeptic

@apooorved introduces Jev as TypeSafe’s first System One model and says it is fast enough for real-time apps, including playing Doom in real time.

Meet Jev (Typesafe's First "System One" model) who knew the next step in AI would be making LLMs obsolete... An AI fast enough for real time applications (inclu…

@njpCoder built a Chrome extension that hides engagement bait, empty hooks, and ads on X. Rules are plain English, and TypeSafe Jev checks each post in half a second.

built a chrome extension that lets me write my own X algorithm. it hides engagement bait, empty hooks and ads before I see them. the rules are plain English. Je…

@YazanO_0 shows a Jev snake demo that survived 321 moves and says it was fast.

#jev it turned out it's good in snake too! fast too! Survived 321 moves! #demo #day_2 #type_safe https://t.co/OFcdy6bsLH

@ali_uraish built an SO-101 pick-and-place simulation. GPT-6 Astra reads overhead and wrist cameras into structured observations, Jev returns typed decisions with probabilities, and a local controller moves the arm and checks limits. The models never drive the motors directly.

Built an SO-101 pick and place simulation using GPT-6 Astra and @typesafeai Jev. Astra reads overhead and wrist camera images and produces structured observatio…

While waiting for Jev access, @SimEdw tried the Doom-style demo with Gemma 4 E2B. It takes 640×480 frames directly, shares the image prefill across batched action queries, and reads controls from logits. They say it sort of works.

The Jev Doom demo by @CompleteSkeptic is really cool, so while I’m waiting for access I tried the same thing with Gemma 4 E2B from @GoogleGemma. Mine takes 640×…

@GeniusGTX describes Jev as an API for structured decisions that can route models, rank search results, and check other AI systems without a chatbot in the loop. Examples are model routing, flagging weak LLM answers, and sorting support tickets. They cite TypeSafe’s report of about 150ms latency and about 100x lower cost and faster responses than LLMs, and say the early-access waitlist is open.

Is Jarvis finally here? TypeSafe just launched Jev, an API for structured decisions that can route models, rank search results and check other AI systems WITHOU…

Quoted @CompleteSkeptic

@keiranhaax told a Hermes agent to test the TypeSafe Jev model on an iOS app they are developing. Hermes recorded the screen and sent it on Telegram. They say more testing will follow.

Told my hermes agent to test the @typesafeai Jev model on an iOS app I’m currently developing for OSS Hermes recorded the screen and sent it to me on telegram I…

Quoted @trycua

@mtropolis_chris demos Jev auto-declining cookie banners and pop-ups on Gymshark, Fashion Nova, and NYPost.

ok so you can also make Jev auto-decline cookie banners and pop ups now here's a demo of it working on different websites (gymshark, fashionnova, nypost) https:…

Quoted @CompleteSkeptic

@TheINAOG cleared Super Mario Bros. 1-1 with Jev plus emulator lookahead for $0.04. Jev sees no pixels: RAM is parsed into structured state, controller macros are simulated, predicted deaths are filtered, and Jev chooses. They resumed from the same state after an API outage, and note that 1-1 is easy for RAM plus search and that skipping vision does not solve perception for general game agents.

Finally I got access to @typesafeai (thank you guys!) Jev + emulator lookahead cleared Super Mario Bros. 1-1 only 0.04$. In this setup, Jev sees no pixels: I pa…

@King_of_Calera posts a slowed-down demo asking whether Jev can run multi-agent swarms that self-heal live productions before clients refresh. The text does not spell out the implementation.

Yes, but can Jev do multi-agent swarms? Enterprise solutions that self-heal live productions before clients can even refresh their browsers? (Demo slowed down f…

@kenonews built a concept demo that turns 144 YouTube uploads from 4 AI channels into a topic map: thumbnails as a wall, titles tagged by topic and hook, then three video ideas. They say the labels are precomputed from titles, not a live Jev run or a speed benchmark.

THE JEV DEMOS SENT ME DOWN A RABBIT HOLE now i have 144 youtube uploads turning into a content map on one screen here’s the concept i built ↓ 144 real videos fr…

@Neel490 made Jev fight OpenJev in a first-person shooter and links the demo at jev-vs-openjev.vercel.app.

Made Jev fight OpenJev in a first person shooter. https://t.co/wSB3psnvwB https://t.co/vwewLu682T

Quoted @ekzhang1

• jev-vs-openjev.vercel.app

@xinqiliu888 built a demo with Jev and GPT-Live-1 that makes music in real time from a description of what the user wants.

Make music in real time, just by describing what you want. A demo I build today with JEV and GPT-Live-1 👇 https://t.co/ndWYSybxiM

Quoted @CompleteSkeptic

@ansonox shows a webpage fight scene driven by Jev. Each block is asked whether it is floor, button, or ignore, then each hunter is asked to chase, wait, flank, or retreat.

Jev can generate a fight scene on your webpage. It asks each block: floor, button, or ignore? Then each hunter asks: chase, wait, flank, or retreat? https://t.c…

@parcadei says to ask Jev whether the available state has enough evidence first, and to classify intent only after the evidence clears a calibrated threshold.

Ask Jev whether the available state has enough evidence first; classify intent only after the evidence clears a calibrated threshold.

@parcadei says to use Jev to find gaps in an active skill from incident reports, then replay past failures as evaluations before promoting an update.

Use Jev to find gaps in an active skill from incident reports, then replay past failures as evaluations before promoting an update.

@parcadei says that after a failed attempt, use Jev to judge whether the retry is meaningfully different; otherwise replan, escalate, or stop.

After a failed attempt, use Jev to judge whether the retry is meaningfully different; otherwise replan, escalate, or stop.

@parcadei says to check task drift at each state transition and to redirect, escalate, or pause the agent when it crosses a threshold.

Check task drift at each state transition and redirect, escalate, or pause the agent when it crosses a threshold.

@parcadei says to use Jev to select the right skill from names and short descriptions before the task reaches the main model.

Use Jev to select the right skill from names and short descriptions before the task reaches the main model.