Jev Computer Use

A 706,048-parameter form-filling scorer from Cua AI: given a UI element and candidate options it returns one probability per option, including check, click, and skip actions.

Python harness for GUI grounding: it divides a screenshot into grid tiles, has a vision model describe each tile neutrally, then asks typesafe-ai/jev to choose among the descriptions. Local confidence and margin gates decide whether to descend or refuse, and it emits only a point and bounding box.

The Show HN for CUA-S1, where Cua's founders ask how many computer-use steps actually need a general-purpose model.

The Cua computer-use platform documents a jev-use recipe where TypeSafe Jev picks bounded action IDs and Cua Driver executes clicks and fills in the browser.

A CLI browser automation tool with optional Jev semantic mode: Surf observes, Jev selects the next action, Surf validates and acts, Jev checks the goal.

A Clash Royale bot that turns a phone screenshot into JSON once a second and asks Jev for a strategy, a card, and a square.

A Python toolkit that puts Jev, Laya, or Cua-S1 at the center of browser, computer-use, robotics, and game agents, with a CLI and an MCP server.

A skeptical computer-use test of Jev with a public test repository, from Builder.io's Steve.

A macOS computer-use loop that OCRs the screen, asks Jev for the next action with typed Choices, and clicks, at about $0.0002 per decision without sending screenshots to a frontier model.

A macOS menu bar app that turns a typed click task into desktop actions: Jev picks from locally detected controls, TipTour executes and validates each action, and a separate Gemini mode handles voice.

An experimental Codex skill and JavaScript runtime that has Jev pick macOS Accessibility elements and actions from text candidates, while local policy gates hold sensitive operations for human confirmation.

A Codex Computer Use plugin that adds bounded Jev decision loops: Codex scopes the task and defines allowed actions, Jev picks an operation and compatible target in one request, and Computer Use performs the click, fill, scroll, or shortcut.

A hands-free macOS voice assistant: whisper.cpp transcribes locally in about 100 ms, then one roughly 250 ms Jev request selects a typed action plus arguments, and code executes it against the Mac.

Voice and typed computer use for macOS: Jev picks the next on-screen action from the Accessibility tree, macOS performs it, and each step is checked for its real effect. No screenshots are sent.

Cua announces an open family of small, specialised System One models for computer use, starting with form filling, and argues narrow tasks deserve narrow models.

A 706,000-parameter model fills a whole form for a fraction of a cent, presented as the specialist case for System One models on narrow tasks.