Jev Atlas
0 中文 Submit a project
Community projectBrowser & agents

jev-browse

By kyrylosyzonenko

A CLI and MCP browser tool pairing Jev decisions with agent-browser actions.

Visit project
jev-browse · demonstration published by the project
Project demonstrationFrom the original project
Overview

Implementation overview

Jev chooses clicks, typing, and completion decisions while agent-browser executes each step. The tool accepts named text values from its caller and exposes an MCP browse tool with a final URL, status, and step log.

Before you start

Typing uses caller-supplied text; page content and supplied values are sent to TypeSafe.

Topics

browserMCPCLIagent-browser

Request reference

Edit a small example and copy a real request before integrating the project.

Try a request

Keep exploring

Browser & agents
001

Jev Ultrafast

Browser Use

A browser loop where Jev selects an operation and page element, with a text model handling field input.

PythonBrowser & agents☆ 4,340
Source
002

Mobile Jev

Droidrun

An Android agent using Jev decisions and Mobilerun, with a live studio, CLI, and execution traces.

JavaScriptBrowser & agents☆ 81
Source
007

A macOS control loop that combines OCR and accessibility data with Jev action selection.

PythonBrowser & agents☆ 192
Source