Jev computer use and browser agents

Malay

@MalayVasa

text-to-screenshot just say what you want and the crosshair goes and gets it. built with jev 1.14 from @typesafeai

Gregor Zunic

@gregpr07

Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓

Marcel Pociot 🧪

@marcelpociot

I built a browser extension with Jev @typesafeai that can hide/collapse posts on X based on natural language. It's so fast that it's not noticeable and insanely cheap...this must be the future of "ad blockers" and content firewalls.

Ronak Malde

@rronak_

Jev + Astra beats the Ender Dragon in Minecraft in 8 minutes 43 seconds! ⏱️ Cost less than $1 ($0.01 Jev, $0.96 Astra) I open sourced the code and explain the harness setup below. This type of movement is only possible with Jev's near instant decisionmaking, and some continually learning skills from Astra.

Cline

@cline

We built a plugin that gives Jev a browser in Cline, and have been blown away by the results. 1. Install it in our new desktop app: Customize > Marketplace > Plugins > search 'jev-browser' 2. Create a Vercel AI Gateway API key, then save it to ~/.cline/plugins/cline-jev-browser.config.json as {"gateway": {"apiKey": "..."}} and restart Cline. 3. Ask any browser task and it will launch Chrome in the background to complete it.

Cua

@trycua

1/ Introducing CUA-S1: a family of System One Models, small, specialized, and built for computer use. Today we're open-sourcing CUA-S1-FORMS, the first in the family: github.com/trycua/cua

idan levin

@0xidanlevin

We just ran Jev on our WebMCP benchmark. The result: basically broke the benchmark. Jev + Mercury 2.5 (a fast, low-cost LLM) using WebMCP solved 100% of the tasks at roughly 112× lower model cost than GPT-6 Astra using computer use with code execution. Compared to Astra using screenshot-based computer use, the model cost was 245× lower (!). We also compared Jev operating the browser with and without WebMCP. We used Browser Use’s open-source Ultrafast, with some improvements to the harness to make it more reliable across the benchmark. Jev’s browser-control accuracy on its own was not amazing - adding WebMCP nearly doubled the number of solved tasks, from 25/49 to 49/49, while reducing model cost by 18% (more on why below). The benchmark and methodology are fully open and reproducible. Full results: https://webmcp.com/benchmark A few words on how the Jev + WebMCP harness works and why this is exciting: Jev receives text as input and a set of discrete options it can choose from. With WebMCP, those options are the tools exposed by the website. At each step, Jev sees the task, the available tools and previous results, then picks what to do next. The limitation is that Jev can’t generate arbitrary text, which you need for tool arguments. For example, it can choose the search_products tool, but it can’t generate the search query itself. So we split the work: Jev picks the tool and Mercury 2.5 generates the arguments if needed. This works well because turns out most of the cognitive load in these tasks is around choosing the right action. The argument generation itself is relatively simple, so we can delegate to a small and very fast model. We used Mercury, which outputs 1,000+ tokens/sec and is very cheap. The result is a pretty simple combination: Jev for tool selection + Mercury for arguments + WebMCP for the interface. It ends up being very reliable, very fast, and very cheap. A few words about Ultrafast and why do we think it underperforms: Without WebMCP, Jev chooses from the page’s controls: which button to click, which field to fill, or which option to select. But choosing a valid button is different from choosing the right next step. The agent still has to navigate menus, understand forms, recover from errors and recognize when the task is actually complete. Our hypothesis is that WebMCP makes the decision space much simpler. Instead of figuring out a sequence of clicks through a website, Jev chooses explicit actions that directly advance the task. @typesafeai itself documents weaker accuracy on questions requiring multiple reasoning steps. WebMCP moves much of that complexity into the website’s tools, leaving Jev with clearer decisions and fewer opportunities to go wrong (in a sense WebMCP "compresses" a sequence of clicks into one tool call). Our modified Ultrafast setup solved 25/49 tasks - that is a result for our particular implementation and benchmark, not a universal limit on Jev or Browser Use. We are open to more harness optimization to get this result to perform better, feel free to directly contribute to the benchmark here: https://github.com/nekuda-ai/WindTunnel Browser-use ultrafast: https://github.com/browser-use/jev-ultrafast

Kyle Jeong

@kylejeong

we built blazing fast computer/browser use with Jev + @Stagehanddev. this task cost $0.001 and executed at near instant speed (in a remote browser btw) the loop: observe the page, send a11y tree as state + actions as questions, Jev decides the next action, then Stagehand executes it.

Milind S

@milindlabs

Okay so Jev can actually do computer use really well Without any screenshots, or LLMs and no Pixels leave my mac I dont even read the Dom elements A local CoreML model segments every button and UI element on screen. On-device OCR reads the labels. That text is all Jev gets. It returns a probability across those elements and tells me the best one to click. Then it clicks, re-runs detection, and decides again. In a loop until the goal is done. ~90ms per decision. Faster than any LLM computer use I've tried. Blazing fast computer use, without any latency @typesafeai is building something really interesting

Niaz Morshed

@niazmorshed_

built `jev-review` @typesafeai it's an experimental, local-first MCP plugin that gives coding agents a score quality feedback loop across different metrics. agents call jev while they work, get scored, make improvements, and repeat the loop try below 👇

Milind S

@milindlabs

got @typesafeai's new model Jev as a chief of staff for bots Jev reads the task, wakes the right teammates off the bench and gives each one the right model It is possible on OpenMausBot as it supports all the LLMs from your existing subscriptions Jev as a decision engine is great

Riley Brown

@rileybrown

Building an agent with model router powered by Jev.

Mormon Negro

@mormonnegro

Acabo de terminar la implementación de @typesafeai + Chromium Headless para que mis agentes puedan navegar por internet a una buena velocidad! En este ejemplo le pido al agente que entre a la página del término "Café" en Wikipedia y navegue por los hipervínculos hasta terminar en la página "Inteligencia artificial" Logra completar la prueba en 20 segundos. Creo que pocos humanos podrían hacerlo en ese tiempo.

Alan Daitch

@AlanDaitch

Claude me integró Jev con Playwright  para buscar usados. Lee unos 26 artículos por minuto y decide qué hacer con cada uno en 406 milisegundos Descartó los que no encajaban con lo que busco, ofertó por los que sí y hasta les mandó un mensaje a los vendedores cuando faltaba algún dato en la publicación. Toda la búsqueda salió USD 0,00085. O sea: con un dólar revisás unas 26.000 publicaciones. Una IA que por fin puede navegar por internet más rápido que nosotros es un game changer

Farouq Aldori

@FarouqAldori

Jev is fun! One-click invoice finder for any website 🧾 - Automatically finds billing pages using @typesafeai's Jev - List/download all invoices with 1 click - Works with Stripe billing portals too - Remembers where invoices live for next time Should I open-source it?

RaZaan

@razaanstha

I built a Chrome extension for agentic browsing using Jev by @typesafeai, fx.sh including AI Gateway by @vercel. Now agents can browse, click, and interact with websites directly in your browser. Cost effective and fassst. Decision-making by Jev.

John Yeo

@johnyeo_

Jev made our Slack agent 2x faster ⚡️ Our agent can be quite slow because it needs to read skills and figure out which tools to call. We used @typesafeai's new model to speed this up by first passing it the prompt and classifying the best skill, tool and params to use before handing it to the agent

Andy Gao

@instantricecook

I built a voice controlled computer-use for my mac using @typesafeai's Jev and it's INSANE how fast it is! I can dictate "open the notes app and create..." and the app opens before I even finish my sentence.

Oskar

@o_kwasniewski

e2e + jev from @typesafeai ⚡ I'm building an open-source framework for running e2e tests with agents. supports web, mobile (and more!) available soon: tester.army/e2e

Marc Köhlbrugge

@marckohlbrugge

building a computer assistant with Jev local whisper listens to everything I say which then gets classified by Jev to determine what actions to take it uses a small Swift app to provide the full accessibility tree to Jev (i.e. tell its what's on my screen, what can be clicked, etc) still super early, but promising and all real-time which Jev was also local though. then it would be completely private

😎Nick 常胜

@isNickMa

Tested TypeSafe’s Jev (no-text, probability-only model) as an AI agent safety monitor. Checking each action first worked well caught most attacks with almost no false blocks, and much faster than Gemini.