News API

Request

curl "https://api.alterlab.io/api/v1/news?q=semiconductors&category=technology&limit=10"

All parameters are optional:

• q: text to match in the headline or snippet.
• category: for example world, technology, business, science, sports or general.
• source: a source id such as bbc, npr, guardian-world or aljazeera.
• language: a language code such as en or de.
• since: an ISO 8601 time; only items published after it are returned.
• limit: 1 to 100 items. Default 20.

Response

{
  "object": "news_list",
  "count": 1,
  "items": [
    {
      "headline": "Chipmakers report record demand",
      "url": "https://example.com/story",
      "source": "bbc",
      "source_name": "BBC News",
      "published_at": "2026-11-11T08:30:00Z",
      "author": null,
      "image": "https://example.com/story.jpg",
      "snippet": "Short summary of the story.",
      "category": "technology",
      "language": "en"
    }
  ],
  "attribution": {"text": "News via AlterLab", "url": "https://alterlab.io/news-api", "required": true},
  "next_action": {"type": "scrape", "endpoint": "POST /api/v1/scrape"},
  "rate_limit": {"tier": "keyless", "limit_per_hour": 60}
}

Items are metadata only: no article text is stored, and every item links to the publisher. To read an article, scrape its url with a normal scrape request; next_action already names that call. An empty result returns next_action.type of refine_query.

Limits and Attribution

Keyless callers get 60 requests per hour per IP. Send an X-API-Key header with a free API key to raise that to 1,000 per hour. Past the limit the API answers 429 with a Retry-After header. Responses are cacheable: keyless responses are shared at the edge for a few minutes. Show the attribution "News via AlterLab" wherever you display the headlines.

MCP Tool

alterlab_news takes the same filters (query, category, source, language, since, limit) and needs no API key. See the News API overview for sources and use cases.