Daily

Daily

Introduction

OpenChrome is an open-source browser automation MCP server that uses your real Chrome browser. It enables AI agents to control Chrome through the Model Context Protocol (MCP), eliminating bot detection and authentication friction.

Key Features

1. Real Chrome Integration

Uses your actual Chrome browser (already logged in to everything)
No more bot detection flags
Invisible to bot detection systems
Authentic sessions across all sites

2. Parallel Execution

Run 20+ parallel Playwright sessions simultaneously
Search across multiple sites at once
Crawl authenticated dashboards in seconds
Single Chrome process at ~300MB (vs 2.5GB for 5 separate browsers)

3. Performance Optimizations

80% fewer LLM calls: Eliminates login navigation and guessing cycles
80x faster wall time: Reduces 20 minutes to 15 seconds
5x cheaper: From ~$2.00 to ~$0.40 per task

4. Intelligent Hint Engine

Watches every tool call across 6 layers
Error recovery, composite suggestions, repetition detection
Learns patterns and provides success guidance
Promotes successful recovery patterns to permanent rules

5. Multi-Account Support

Isolated sessions for each account
Check orders across personal and business accounts simultaneously
Perfect for competitive intelligence (price comparison, etc.)

6. Rich Tool Set (47 tools)

Navigation & Interaction:

navigate, click_element, fill_form, wait_and_click, find, computer

Read & Extract:

read_page, page_content, javascript_tool, selector_query, xpath_query

Environment:

emulate_device, geolocation, user_agent, network

Storage & Debug:

cookies, storage, console_capture, performance_metrics, request_intercept

Parallel Workflows:

workflow_init, workflow_collect, worker_create, batch_execute Memory:
memory_record, memory_query, memory_validate

7. Three read_page Modes

ax (default): Accessibility tree with ref_N IDs
baseline: Screen readers, semantic analysis
dom: Compact DOM with stable backendNodeId IDs (~5-10x fewer tokens)
css: CSS diagnostic info for debugging
minimal: Minimal mode for debugging styles

8. MCP Protocol Support

Native MCP server implementation
Works with Claude Code, VS Code, Cursor, Windsurf, Copilot
Auto-configuration with oc setup
CLI commands: oc, oc serve, oc doctor

Quick Start

One-Command Setup

npx openchrome-mcp setup

Restart Claude Code, then use oc commands.

Manual Configuration

Claude Code:

claude mcp add openchrome -- npx -y openchrome-mcp@latest serve --auto-launch

VS Code / Copilot (.vscode/mcp.json):

{
  "servers": {
    "openchrome": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "openchrome-mcp@latest", "serve", "--auto-launch"]
    }
  }
}

Use Cases

Competitive intelligence: Compare prices across multiple sites simultaneously
Multi-account management: Check orders across personal and business accounts
Parallel monitoring: Screenshot AWS billing, GCP console, Stripe at once
Debug production UIs: Test with real user sessions
Web scraping: Large-scale data collection

Benchmarks

OpenChrome provides comprehensive benchmarks to measure performance:

Token efficiency and parallel performance
Multi-step interaction, batch JS execution, compiled plan execution
Streaming collection, init overhead, fault tolerance
Scalability curves with real-world tests

Why Choose OpenChrome?

If you need browser automation with real Chrome sessions, parallel execution, and intelligent error recovery, OpenChrome is a powerful choice. It integrates seamlessly with Claude Code and other MCP clients, giving your AI agent browser superpowers across Telegram, Discord, or any chat platform.

Try it out and share your experience!

#MCP #BrowserAutomation #Playwright #Chrome #AI #OpenSource #ClaudeCode #DeveloperTools

OpenChrome: Open-Source Browser Automation MCP Server