News MCP server
Connect Claude Desktop, Cursor, or any MCP client to live, already-scored news data. Your agent asks for "breaking sanctions news from Russia, urgency 7+" — and gets back classified, de-duplicated articles. No NLP pipeline to build.
What is this?
The Model Context Protocol lets AI agents call external tools. Our MCP server wraps the NewsAgent Data API and exposes five news tools your agent can use natively — every article it returns already carries an urgency score (0–10), political-lean label, topic, country and audience tags.
The five tools
| Tool | What it does |
|---|---|
get_feed | Filtered feed — country, topic, language, political lean, audience, min urgency, date range |
search_news | Full-text keyword search across the archive |
get_breaking | Recent high-urgency news (≥ 7 by default) |
coverage_stats | Live totals — articles, sources, countries, languages (no key) |
list_sources | Source catalog metadata (Standard tier) |
Setup (2 minutes)
1. Get a free API key — sign up here, no card.
2. Download the server and install deps:
pip install mcp httpx
3. Add it to your client. Claude Desktop (claude_desktop_config.json):
{
"mcpServers": {
"newsagent": {
"command": "python",
"args": ["/absolute/path/to/newsagent_mcp.py"],
"env": { "NEWSAGENT_API_KEY": "your_key_here" }
}
}
}
Restart the client — the five tools appear automatically. Cursor and Cline use the same command + env.
Try saying
- "Get breaking sanctions news from Russia scored 7 or higher."
- "Search for 'central bank rate' in the last 14 days, English only."
- "Compare how state vs opposition sources covered Ukraine this week."
- "How many countries and sources do you cover right now?"
The agent picks the right tool, applies the filters, and gets back pre-scored, classified results.
Grab a free API key and wire up the MCP server in two minutes.
Get a free API key
See also: API documentation · News API for LLM/RAG · News by topic