Report Abuse

Basic Information

This repository provides a desktop application that makes websites accessible to AI agents by packaging and extending the browser-use project and web-ui into a user-facing Electron/Gradio UI. The app is intended to run locally and to let users interact with a browser agent through a graphical interface rather than only programmatic APIs. It integrates multiple large language model backends so agents can be paired with different LLM providers while using the real Chrome browser on the host machine. The README documents a simple developer get-started flow using git clone, npm install and vite dev, and notes planned downloadable installers for macOS, Windows and Linux. The overall aim is to simplify running, testing and interacting with browser-based agents while leveraging existing browser sessions and credentials.

Links

App Details

Features
A desktop UI built on Electron and Gradio that exposes most browser-use functionality for easy interaction with a browser agent. Expanded LLM support with integrations listed in the README including Google, OpenAI, Azure OpenAI, Anthropic, DeepSeek and Ollama, with plans to add more. Custom browser support that uses the user"s existing Google Chrome profile to avoid re-login and authentication challenges. High-definition screen recording support is indicated for capturing agent-driven sessions. Simple development startup instructions are provided (git clone, npm install, vite dev). Downloadable builds for major platforms are noted as coming soon. The project builds on and interoperates with the browser-use and web-ui projects rather than reinventing core agent controls.
Use Cases
The app reduces friction when building and testing web automation agents by letting them run against an existing Chrome session so users do not need to re-authenticate sites, which accelerates development and troubleshooting. Multi-LLM support enables experimentation with different model providers and comparisons without rearchitecting the client. The Electron/Gradio UI makes agent controls accessible to less technical users and speeds prototyping by exposing browser-use features in a graphical environment. HD screen recording helps with reproducibility and demonstration of agent behaviors. The project integrates with the broader browser-use ecosystem which aids developers already using those components and simplifies setup via the provided install and local dev commands.

Please fill the required fields*