# google-news [Health: Active]

**Category:** 🔎 Search & Data Extraction  
**Repository:** https://github.com/johnisanerd/Apify-Google-News-Scraper  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/google-news-2

## Description
Google News headlines, sources, and links via the Apify Google News API, hosted MCP.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "google-news": {
    "url": "https://mcp.apify.com/?tools=actors,docs,johnvc/GoogleNewsAPI"
  }
}
```

## Documentation

## What the google-news MCP server does

The google-news MCP server exposes an Apify Google News Actor through a hosted MCP endpoint. It accepts a required search query and returns Google News results in JSON. Results are organized one item per fetched page, with a `news_results` array containing each article's title, link, source, snippet, and ranking position.

Each page also includes the submitted search parameters, the displayed query, a timestamp, the page number, and metadata such as total results, article count, processed pages, and whether the pagination limit was reached. This makes the response usable in agent workflows that need both article data and context about how the search ran.

## How it works

The hosted endpoint is backed by the Apify Actor identified as `johnvc/GoogleNewsAPI`. Connect an MCP client to the preconfigured URL:

```text
https://mcp.apify.com/?tools=actors,docs,johnvc/GoogleNewsAPI
```

Authentication can use Apify OAuth in a browser or an Apify API token. The same token can also be used by the repository's Python example. The Actor searches Google News and returns one result item for each requested page, with roughly ten articles per page according to the documented input behavior.

The query supports a Google domain, country code, language code, language restriction, and a location string. `safe` controls safe search, `nfpr` can prevent auto-corrected results, and `filter` can remove duplicates. `max_pages` limits the search depth; setting it to `0` requests unlimited pages, while each fetched page is billed separately by Apify.

## Setup and configuration

For local use, the repository requires Python 3.11 or newer, an Apify account, and an Apify API key. Install dependencies with UV, create a `.env` file from `.env.example`, and set `APIFY_API_TOKEN`. The documented example runs with:

```bash
uv run python google-news-scraper.py
```

Alternatively, export `APIFY_API_TOKEN` in the shell before running the command. The optional `output_file` input can specify a filename for saving results.

For MCP use, add the hosted Apify endpoint to a compatible client. Claude Desktop can use `mcp-remote` in its MCP configuration, and Claude Code can register the endpoint with an HTTP transport. OAuth or a bearer token is required when the tool is first used.

## Tools and capabilities

The google-news MCP server supports a single Google News search Actor with these main capabilities:

- Search Google News with a required `q` string.
- Target results by location, Google domain, country, or language.
- Apply safe-search and duplicate-filtering options.
- Preserve exact query intent by controlling auto-correction.
- Fetch a selected number of result pages.
- Return consistent article fields and search metadata as JSON.

## Limitations and notes

The google-news MCP server depends on Apify authentication and billing. The documented pricing model includes a per-run setup charge and a per-page charge, so increasing `max_pages` can increase usage costs. Results come from Google News' aggregated index; this tool does not provide a separate crawler for individual publishers.

The material documents Claude Desktop, Claude Code, and Cursor as client options or examples. It does not document article full text, feeds, alerts, sentiment analysis, or historical archives. The MCP endpoint is hosted through Apify, while the repository also provides a local Python execution path.

_Full upstream README: https://allmcps.com/mcp/google-news-2/readme_