What makes this different from standard browser tools?

This skill provides truly isolated sessions for multiple agents, each with its own cookies and storage, plus a high-speed screencast API for near-instant visual updates.

Can I use my existing Chrome logins with this skill?

Yes, by using the --profile flag when starting the server, it can copy your default Chrome profile to provide authenticated access to websites.

How fast are the visual updates?

The server provides cached frames from a continuous screencast in approximately 15ms, which is significantly faster than triggering fresh renders for every screenshot.

Does it support headless mode?

Yes, it is designed to run in headless mode by default, making it ideal for automated agent tasks and environments without a display.

How are sessions cleaned up?

Sessions can be manually destroyed via the API once a task is complete, and the server includes an automatic 5-minute idle timeout for resource management.

Multi-Session Browser Server

Name: Multi-Session Browser Server
Author: tavva

bytavva

Browser Automation

Provides isolated browser contexts and high-performance screencasting for multi-agent web automation, visual debugging, and web scraping.

About

This skill enables Claude to interact with web pages through a high-performance HTTP server that manages multiple isolated Chrome sessions simultaneously. By providing separate cookies and local storage for each session alongside ultra-fast screencast frame retrieval (~15ms), it allows developers to verify frontend changes, debug layout issues, and perform complex web scraping tasks without session interference. The server effectively turns a headless Chrome instance into a scalable automation hub for AI agents.

Key Features

High-speed screencast frame retrieval (~15ms) for rapid visual iteration
Isolated browser contexts with unique cookies and localStorage per session
High-resolution PNG screenshots for OCR and precise visual inspections
Parallel multi-agent support for simultaneous browsing tasks
0 GitHub stars
Automatic session lifecycle management and idle timeout cleanup

Use Cases

Verifying frontend UI changes and debugging layout issues in real-time
Scraping data from multiple authenticated web pages simultaneously
Testing complex web features across isolated user sessions for multi-role workflows

About

Key Features

High-speed screencast frame retrieval (~15ms) for rapid visual iteration
Isolated browser contexts with unique cookies and localStorage per session
High-resolution PNG screenshots for OCR and precise visual inspections
Parallel multi-agent support for simultaneous browsing tasks
0 GitHub stars
Automatic session lifecycle management and idle timeout cleanup

Use Cases

Verifying frontend UI changes and debugging layout issues in real-time
Scraping data from multiple authenticated web pages simultaneously
Testing complex web features across isolated user sessions for multi-role workflows