Bruster: A Faster-Starting, Lighter Browser for AI Agents

Description and use

Bruster is an open-source browser automation tool written in Rust, designed for AI agents and scripts. It uses native webviews and exposes MCP and JSON lines interfaces. A published Linux benchmark shows faster startup (219 ms vs. 799 ms) and lower memory (256 MiB vs. 523 MiB PSS) than agent-browser 0.40.0.

Problem

AI agents and scripts need a fast-starting, low-memory browser automation tool that works with native webviews and MCP clients.

Outcome

Bruster achieves faster startup (219 ms vs. 799 ms) and lower memory (256 MiB vs. 523 MiB PSS) than agent-browser 0.40.0 in a published Linux benchmark on Ubuntu 24.04.

Workflow

• Install Bruster via install.sh on macOS or Linux
• Register the MCP server with Claude Code or another MCP-compatible client
• Use Bruster to navigate websites, inspect page content, click elements, fill forms, submit forms, capture screenshots, and manage tabs/sessions

Evidence

The goal is simple: fast startup, lower memory usage, and a straightforward way to control a real browser.

Source

Here’s one published Linux benchmark comparing Bruster with agent-browser 0.40.0 using full Chrome: Launch to first loaded page: 219 ms vs. 799 ms; Total idle memory, PSS: 256 MiB vs. 523 MiB.

Source

Linux requires WebKitGTK 4.1, plus xvfb and xauth when running without a desktop display.

Source

Current limitations include synthetic input events and no cross-origin iframe access.

Source