← back to Norma
agents/datasource-agent/lib/newspaper-registry.js
89 lines
/**
* Newspaper Registry — US and international newspaper RSS feed URLs.
*
* Used by the scrape-frontpages skill to discover daily headlines.
* Each entry includes name, RSS URL, country code, region, and source type.
*/
const US_NEWSPAPERS = [
{ name: 'The New York Times', url: 'https://rss.nytimes.com/services/xml/rss/nyt/HomePage.xml', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'The Washington Post', url: 'https://feeds.washingtonpost.com/rss/national', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'Wall Street Journal', url: 'https://feeds.a.dj.com/rss/RSSWorldNews.xml', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'Los Angeles Times', url: 'https://www.latimes.com/world-nation/rss2.0.xml', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'Chicago Tribune', url: 'https://www.chicagotribune.com/arcio/rss/category/nation-world/', country: 'US', region: 'Midwest', source_type: 'domestic' },
{ name: 'Boston Globe', url: 'https://www.bostonglobe.com/arc/outboundfeeds/rss/homepage/', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'USA Today', url: 'http://rssfeeds.usatoday.com/UsatodaycomNation-TopStories', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'NPR', url: 'https://feeds.npr.org/1001/rss.xml', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'PBS NewsHour', url: 'https://www.pbs.org/newshour/feeds/rss/headlines', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Associated Press', url: 'https://rsshub.app/apnews/topics/apf-topnews', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Reuters', url: 'https://www.reutersagency.com/feed/', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'CNN', url: 'http://rss.cnn.com/rss/cnn_topstories.rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'MSNBC', url: 'https://www.msnbc.com/feeds/latest', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Fox News', url: 'https://moxie.foxnews.com/google-publisher/latest.xml', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'CBS News', url: 'https://www.cbsnews.com/latest/rss/main', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'ABC News', url: 'https://abcnews.go.com/abcnews/topstories', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'NBC News', url: 'https://feeds.nbcnews.com/nbcnews/public/news', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Bloomberg', url: 'https://feeds.bloomberg.com/politics/news.rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Politico', url: 'https://rss.politico.com/politics-news.xml', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'The Hill', url: 'https://thehill.com/feed/', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'HuffPost', url: 'https://www.huffpost.com/section/front-page/feed', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'ProPublica', url: 'https://feeds.propublica.org/propublica/main', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'The Atlantic', url: 'https://www.theatlantic.com/feed/all/', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Vox', url: 'https://www.vox.com/rss/index.xml', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Axios', url: 'https://api.axios.com/feed/', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'The Guardian US', url: 'https://www.theguardian.com/us-news/rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'New York Post', url: 'https://nypost.com/feed/', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'The Dallas Morning News', url: 'https://www.dallasnews.com/arcio/rss/', country: 'US', region: 'South', source_type: 'domestic' },
{ name: 'Houston Chronicle', url: 'https://www.houstonchronicle.com/rss/feed/Breaking-News-702.php', country: 'US', region: 'South', source_type: 'domestic' },
{ name: 'Miami Herald', url: 'https://www.miamiherald.com/latest-news/index.rss', country: 'US', region: 'South', source_type: 'domestic' },
{ name: 'San Francisco Chronicle', url: 'https://www.sfchronicle.com/rss/feed/News-702.php', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'The Seattle Times', url: 'https://www.seattletimes.com/feed/', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'The Denver Post', url: 'https://www.denverpost.com/feed/', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'The Philadelphia Inquirer', url: 'https://www.inquirer.com/arcio/rss/category/news/', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'Minneapolis Star Tribune', url: 'https://www.startribune.com/rss.xml', country: 'US', region: 'Midwest', source_type: 'domestic' },
{ name: 'The Detroit News', url: 'https://www.detroitnews.com/rss/', country: 'US', region: 'Midwest', source_type: 'domestic' },
{ name: 'Atlanta Journal-Constitution', url: 'https://www.ajc.com/arcio/rss/category/news/', country: 'US', region: 'South', source_type: 'domestic' },
{ name: 'The Oregonian', url: 'https://www.oregonlive.com/arc/outboundfeeds/rss/category/news/', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'St. Louis Post-Dispatch', url: 'https://www.stltoday.com/search/?f=rss&t=article&l=15', country: 'US', region: 'Midwest', source_type: 'domestic' },
{ name: 'Tampa Bay Times', url: 'https://www.tampabay.com/arcio/rss/category/news/', country: 'US', region: 'South', source_type: 'domestic' },
{ name: 'The Arizona Republic', url: 'https://rssfeeds.azcentral.com/phoenix/topstories', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'Pittsburgh Post-Gazette', url: 'https://www.post-gazette.com/rss/nation', country: 'US', region: 'Northeast', source_type: 'domestic' },
{ name: 'The Sacramento Bee', url: 'https://www.sacbee.com/latest-news/index.rss', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'The San Diego Union-Tribune', url: 'https://www.sandiegouniontribune.com/arcio/rss/', country: 'US', region: 'West', source_type: 'domestic' },
{ name: 'The Intercept', url: 'https://theintercept.com/feed/?rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Slate', url: 'https://slate.com/feeds/all.rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'The Daily Beast', url: 'https://feeds.thedailybeast.com/rss/articles', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'Business Insider', url: 'https://www.businessinsider.com/rss', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'The Economist US', url: 'https://www.economist.com/united-states/rss.xml', country: 'US', region: 'National', source_type: 'domestic' },
{ name: 'MarketWatch', url: 'https://feeds.marketwatch.com/marketwatch/topstories/', country: 'US', region: 'National', source_type: 'domestic' },
];
const INTERNATIONAL_NEWSPAPERS = [
{ name: 'BBC News', url: 'http://feeds.bbci.co.uk/news/rss.xml', country: 'GB', region: 'Europe', source_type: 'international' },
{ name: 'The Guardian UK', url: 'https://www.theguardian.com/world/rss', country: 'GB', region: 'Europe', source_type: 'international' },
{ name: 'Financial Times', url: 'https://www.ft.com/rss/home', country: 'GB', region: 'Europe', source_type: 'international' },
{ name: 'The Telegraph', url: 'https://www.telegraph.co.uk/rss.xml', country: 'GB', region: 'Europe', source_type: 'international' },
{ name: 'Der Spiegel', url: 'https://www.spiegel.de/international/index.rss', country: 'DE', region: 'Europe', source_type: 'international' },
{ name: 'Le Monde', url: 'https://www.lemonde.fr/en/rss/une.xml', country: 'FR', region: 'Europe', source_type: 'international' },
{ name: 'El Pais', url: 'https://feeds.elpais.com/mrss-s/pages/ep/site/english.elpais.com/portada', country: 'ES', region: 'Europe', source_type: 'international' },
{ name: 'Corriere della Sera', url: 'https://xml2.corriereobjects.it/rss/homepage.xml', country: 'IT', region: 'Europe', source_type: 'international' },
{ name: 'The Globe and Mail', url: 'https://www.theglobeandmail.com/arc/outboundfeeds/rss/category/canada/', country: 'CA', region: 'North America', source_type: 'international' },
{ name: 'Sydney Morning Herald', url: 'https://www.smh.com.au/rss/feed.xml', country: 'AU', region: 'Oceania', source_type: 'international' },
{ name: 'The Japan Times', url: 'https://www.japantimes.co.jp/feed/', country: 'JP', region: 'Asia', source_type: 'international' },
{ name: 'South China Morning Post', url: 'https://www.scmp.com/rss/91/feed', country: 'HK', region: 'Asia', source_type: 'international' },
{ name: 'Al Jazeera', url: 'https://www.aljazeera.com/xml/rss/all.xml', country: 'QA', region: 'Middle East', source_type: 'international' },
{ name: 'Times of India', url: 'https://timesofindia.indiatimes.com/rssfeedstopstories.cms', country: 'IN', region: 'Asia', source_type: 'international' },
{ name: 'The Irish Times', url: 'https://www.irishtimes.com/cmlink/news-1.1319192', country: 'IE', region: 'Europe', source_type: 'international' },
{ name: 'New Zealand Herald', url: 'https://www.nzherald.co.nz/arc/outboundfeeds/rss/section/nz/', country: 'NZ', region: 'Oceania', source_type: 'international' },
{ name: 'The Straits Times', url: 'https://www.straitstimes.com/news/world/rss.xml', country: 'SG', region: 'Asia', source_type: 'international' },
{ name: 'Dawn', url: 'https://www.dawn.com/feeds/home', country: 'PK', region: 'Asia', source_type: 'international' },
{ name: 'Haaretz', url: 'https://www.haaretz.com/cmlink/1.628752', country: 'IL', region: 'Middle East', source_type: 'international' },
{ name: 'The Toronto Star', url: 'https://www.thestar.com/search/?f=rss&t=article&l=15', country: 'CA', region: 'North America', source_type: 'international' },
];
module.exports = {
US_NEWSPAPERS,
INTERNATIONAL_NEWSPAPERS,
ALL_NEWSPAPERS: [...US_NEWSPAPERS, ...INTERNATIONAL_NEWSPAPERS],
};