Chrome History & Bookmarks

You run a Model Context Protocol (MCP) server that gives AI assistants access to Chrome history and bookmarks data for powerful, private analysis. This server reads your local Chrome data in read-only mode, exposes practical tools to search and summarize history, and supports integration with MCP clients to drive tailored insights directly from your browser activity.

How to use

You connect an MCP client to the server to perform common tasks like searching your history, viewing recent activity, and getting statistics about your Chrome data. Start by launching the local MCP server, then configure your client to point at the local runtime. Use the provided tool names to request specific actions such as search_history, get_recent_browsing, and get_history_stats. All processing happens locally on your machine with read-only access to Chrome data.

How to install

Prerequisites: You need Node.js 16 or newer and Chrome/Chromium installed (Chrome should be run at least once to establish profile data). You also need an MCP-compatible client to connect to the server.

Step 1: Clone the project repository locally.

git clone https://github.com/your-username/chrome-history-mcp-server.git
cd chrome-history-mcp-server

Step 2: Install dependencies with your package manager.

npm install

Step 3: Run the MCP server locally.

node src/mcp-server.js

Configuration and usage notes

The server is designed to locate Chrome profile data and expose tools in a cross-platform manner. It reads Chrome's History database in read-only mode and serves results to MCP clients without modifying your data.

If you want to run the server from a client configuration, you can reference the following local runtime example that you would place in your MCP client config:

{
  "mcpServers": {
    "chrome_history": {
      "command": "node",
      "args": ["src/mcp-server.js"],
      "env": {}
    }
  }
}

Security & privacy

Read-only access ensures Chrome data cannot be modified. All processing happens locally on your machine with no outbound network requests, preserving privacy. The server uses SQLite read-only mode to interact with the History database and avoid any data corruption.

Troubleshooting

If you encounter a "Chrome installation not found" error, verify that Chrome is installed and has been run at least once. If a custom profile is used, ensure the path is accessible and Chrome has created a profile directory.

If you see a "Database query failed" error, make sure Chrome is closed while the server runs to avoid database locks, and confirm the Chrome profile permissions allow read access.

For date-range results, use get_history_stats to determine the available span and ensure you format dates as YYYY-MM-DD.

Notes and future plans

Current capabilities cover history search, recent activity, and basic statistics. Bookmarks analysis, advanced browsing pattern insights, and multiple export formats are planned for future releases.

Available tools

search_history

Search through Chrome browsing history with optional date filtering and keyword-based queries.

get_recent_browsing

Retrieve browsing activity from the most recent period with details and timestamps.

get_history_stats

Produce statistics about the history database, including date range and total entries.