Google News MCP Agent

An MCP (Model Context Protocol) server that fetches news articles from Google News, vectorizes them locally using ChromaDB, and allows for semantic search. This tool enables AI agents (like Claude) to stay updated with specific topics and query stored knowledge effectively.

Features

• Ingest News: Fetch headlines and summaries from Google News by topic.
• Local Vector Store: Automatically embeds and stores articles in a local ChromaDB instance (persisted in chroma_db/).
• Semantic Search: Search through the ingested news using natural language queries to find relevant information.
• Privacy First: Runs entirely locally (excluding the initial news fetch). No API keys required for embeddings (uses sentence-transformers).

Prerequisites

• Python 3.10 or higher
• uv (recommended) or pip

Installation

Using uv (Recommended)

1. Clone the repository:

   git clone https://github.com/Tatsuya50/google-news-mcp.git
   cd google-news-mcp
   

2. Install dependencies:

   uv sync
   

Using pip

1. Clone the repository and navigate to the directory.
2. Install the required packages:

   pip install -r requirements.txt
   

Configuration (Claude Desktop)

To use this with Claude Desktop, add the following configuration to your MCP config file (typically ~/AppData/Roaming/Claude/claude_desktop_config.json on Windows):

{
  "mcpServers": {
    "google-news": {
      "command": "uv",
      "args": [
        "--directory",
        "C:\\Users\\YOUR_USERNAME\\path\\to\\google-news-mcp",
        "run",
        "python",
        "mcp_server.py"
      ]
    }
  }
}

Note: Replace C:\\Users\\YOUR_USERNAME\\path\\to\\google-news-mcp with the actual absolute path to this repository.

Tools

ingest_news

Fetches and indexes news articles.

• topic: The topic to search for (e.g., "Generative AI", "Stock Market").
• max_results: (Optional) Number of articles to fetch (default: 5).

search_news

Searches the stored local database.

• query: The question or topic to search for (e.g., "What are the latest AI trends?").
• n_results: (Optional) Number of results to return (default: 3).

Development

Run the server locally for testing:

uv run python mcp_server.py

Inspect the database contents:

uv run python inspect_db.py