Simple Console

You have a minimal MCP Server designed to monitor browser Console Logs and run JavaScript in-page. It’s intentionally lightweight to keep token context small while you debug in real time. With four focused tools, you can list tabs, capture console output, navigate pages, and execute JavaScript in the page context.

How to use

You interact with the MCP server from your MCP client (such as Claude Desktop or Claude Code). The server exposes four tools that let you list browser targets, read console logs, navigate or reload pages, and run JavaScript in the page. Use these to iteratively inspect errors, warnings, and the page state while debugging.

How to install

Prerequisites you need before installation:

• Node.js 18+ installed on your machine.

• A working MCP client that supports stdio MCP servers (Claude Code/Desktop is compatible).

Choose one of the installation methods below.

# Option 1: use npx through an MCP client configuration
# Example command to add the server via Claude Code/Desktop
claude mcp add simple-console -- npx -y simple-console-mcp

# Alternative inline JSON configuration for Claude/Desktop-compatible setup
{
  "mcpServers": {
    "simple-console": {
      "command": "npx",
      "args": ["-y", "simple-console-mcp"]
    }
  }
}

Additional installation options

If you prefer using a GitHub URL for the MCP server, you can point the client at the repository directly.

claude mcp add simple-console -- npx -y github:tznthou/simple-console-mcp

Notes about starting and running

The server is designed to auto-detect and work with Chrome's CDP interface. When you start it, it will either connect to an existing Chrome instance that has debugging enabled or automatically launch a debug Chrome with a separate user profile. If a regular Chrome window blocks the connection, you’ll see a clear error and you can close that instance to proceed.

If you need to run Chrome manually, start it with remote debugging on port 9222 as a fallback.

Security and notes

Key behavior to know about the internal log handling and navigation: the server maintains a log cache per target with a limit of 500 entries. Navigating a page clears that target’s log cache. Logs are pulled by your client on demand; the MCP protocol is request-response-based and does not push updates automatically.

Tools overview

The server provides four core tools you can call from your MCP client:

• list_targets — List browser targets (pages, service workers, etc.)
• get_console_logs — Read console output from a target (start monitoring on first call)
• navigate — Navigate to a URL or reload the page
• execute_js — Run JavaScript in the page context (with safety limits)

Expected inputs and behavior

When using the tools, you supply parameters to control which target you work with and how much data you retrieve. Defaults are chosen to be sensible for typical debugging workflows.

Examples of typical usage

# List available targets
list_targets

# Read console logs from the first target
get_console_logs --targetIndex 0 --maxLines 50 --filter all

# Navigate to a URL
navigate --url https://example.com --targetIndex 0

# Execute a simple script
execute_js --code "document.title" --targetIndex 0

Available tools

list_targets

List all available browser targets (pages, Service Workers, etc.).

get_console_logs

Get Console output from a specific target. Starts monitoring on first call.

navigate

Navigate to a URL or reload the current page.

execute_js

Execute JavaScript code in the page context with safety limits and timeouts.