Google News Scraper

Scrape and monitor Google News by keyword, date, and publisher. Detect only new stories, filter noise, and optionally add sentiment and entities. No API key.

Drop in your search terms and get matching Google News articles with title, publisher, publisher domain, publish date, snippet, and link. Add rolling or explicit date windows, exact phrases, excluded terms, publisher allowlists/blocklists, monitoring state, and lightweight analysis when you need them.

⚡ Works out of the box — reads Google News' public RSS feeds. No API key, no signup, no proxy needed for normal volumes.

---

What it does

For each search query, the Actor reads the official Google News search RSS feed and extracts structured article data — articles matching any query you provide (e.g. openai, interest rates, a brand name, a ticker).

Output is one row per article, deduplicated across all queries and sorted newest-first. In monitoring mode, later runs emit only articles that have not been seen for the same monitor key and query configuration.

---

Who it's for

• PR & comms teams — monitor brand, executives, and competitor coverage.
• Market researchers & analysts — track an industry, ticker, or theme over time.
• Journalists — sweep a beat across many sources at once.
• Content & SEO teams — find trending angles and source material.
• Builders — pipe a structured news feed into dashboards, alerts, or LLM workflows.

---

Popular monitoring use cases

• Brand monitoring - track mentions of your company, products, founders, and executives.
• Competitor news tracking - monitor launches, funding, hiring, lawsuits, partnerships, and pricing changes.
• Industry trend monitoring - follow themes like AI regulation, climate tech, fintech, or healthcare policy.
• Stock and market news feeds - collect recent Google News results for tickers, companies, and sectors.
• PR and crisis alerts - schedule searches around sensitive topics and export fresh coverage.
• LLM news pipelines - feed headlines, snippets, sources, and dates into dashboards or AI workflows.

---

Input

Field	Type	Default	Description
queries	string[]	—	Search terms
queriesText	string	—	Queries as text (one per line / comma-separated)
exactPhrase	string	""	Require an exact phrase
excludeKeywords	string[]	[]	Exclude words or phrases
includeDomains	string[]	[]	Only keep these publisher domains
excludeDomains	string[]	[]	Remove these publisher domains
postedWithinDays	integer	—	Rolling 1–365 day window
dateFrom / dateTo	string	—	Inclusive YYYY-MM-DD date range
geo	string	"US"	Country code (gl)
language	string	"en"	Language code (hl)
maxItemsPerQuery	integer	50	Cap per query/topic (max 100)
maxQueries	integer	100	Max search queries
dedupe	boolean	true	Remove duplicate articles across queries
enableAnalysis	boolean	false	Add lightweight sentiment, score, and headline entities
monitoringMode	boolean	false	Persist state and emit only unseen articles
monitorKey	string	"default"	Identifier for an independent monitor
monitoringInitialRun	string	"emit"	emit current stories or create a silent baseline
requestDelayMs	integer	500	Delay between feed requests
maxConcurrency	integer	3	Parallel feed requests
proxyConfiguration	object	—	Optional, for high volume

At least one of queries / queriesText is required.

Example input

{
  "queries": ["openai", "climate tech", "interest rates"],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50
}

More in examples/.

Advanced filtering

The Actor applies supported Google News query operators and then validates publisher, keyword, and date filters against the parsed rows. Use postedWithinDays for rolling alerts or dateFrom / dateTo for a fixed range, not both.

{
  "queries": ["artificial intelligence"],
  "exactPhrase": "product launch",
  "excludeKeywords": ["rumor", "opinion"],
  "includeDomains": ["reuters.com", "techcrunch.com", "theverge.com"],
  "postedWithinDays": 14,
  "enableAnalysis": true,
  "geo": "US",
  "language": "en"
}

Only-new monitoring

Enable monitoringMode on a saved task and schedule it. The first run can output the current snapshot (emit) or silently establish a baseline (baseline). Later runs with the same monitor key and query configuration output only unseen articles.

{
  "queries": ["OpenAI", "Anthropic", "Google Gemini"],
  "postedWithinDays": 7,
  "monitoringMode": true,
  "monitorKey": "ai-competitor-alerts",
  "monitoringInitialRun": "baseline",
  "dedupe": true
}

---

Ready-to-run task ideas

Brand News Monitoring

{
  "queriesText": "your brand name\nyour ceo name\nyour main competitor\nyour product line",
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 100,
  "dedupe": true,
  "maxConcurrency": 2
}

Competitor Launch Tracker

{
  "queries": [
    "competitor funding",
    "competitor product launch",
    "competitor partnership",
    "competitor acquisition"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50,
  "dedupe": true
}

AI Industry News Feed

{
  "queries": [
    "artificial intelligence regulation",
    "generative AI startups",
    "AI search engines",
    "large language models"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 75,
  "dedupe": true
}

Stock Market News Monitor

{
  "queries": [
    "Nvidia stock news",
    "Tesla stock news",
    "Microsoft earnings",
    "Federal Reserve interest rates"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50,
  "dedupe": true
}

Crisis & PR Alert Feed

{
  "queries": [
    "\"your brand\" lawsuit",
    "\"your brand\" outage",
    "\"your brand\" data breach",
    "\"your brand\" controversy"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 100,
  "dedupe": true,
  "maxConcurrency": 2
}

Startup Funding News

{
  "queries": [
    "startup funding round",
    "venture capital investment",
    "series A funding",
    "startup acquisition"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 75,
  "dedupe": true
}

Cybersecurity Breach Monitor

{
  "queries": [
    "data breach",
    "ransomware attack",
    "cybersecurity incident",
    "security vulnerability"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 75,
  "dedupe": true
}

Healthcare Policy News

{
  "queries": [
    "healthcare policy",
    "Medicare news",
    "FDA approval",
    "health insurance regulation"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50,
  "dedupe": true
}

Crypto Market News

{
  "queries": [
    "Bitcoin ETF news",
    "Ethereum regulation",
    "crypto market news",
    "stablecoin legislation"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50,
  "dedupe": true
}

Real Estate Market Monitor

{
  "queries": [
    "mortgage rates",
    "housing market news",
    "commercial real estate",
    "home prices"
  ],
  "geo": "US",
  "language": "en",
  "maxItemsPerQuery": 50,
  "dedupe": true
}

---

Output

One dataset row per article:

{
  "query": "openai",
  "queryType": "search",
  "title": "OpenAI ships new model & tools",
  "googleNewsUrl": "https://news.google.com/rss/articles/ABC123?oc=5",
  "source": "The Verge",
  "sourceUrl": "https://www.theverge.com",
  "sourceDomain": "theverge.com",
  "publishedAt": "2026-06-09T14:30:00.000Z",
  "snippet": "OpenAI ships new model The Verge",
  "guid": "ABC123",
  "position": 1,
  "geo": "US",
  "language": "en",
  "fetchedAt": "2026-06-24T12:00:00.000Z",
  "appliedQuery": "openai when:7d",
  "sentiment": "positive",
  "sentimentScore": 0.67,
  "entities": ["OpenAI"],
  "isNew": true
}

Key-value store outputs

• HEADLINES_BY_SOURCE — article titles grouped by publisher
• SUMMARY — fetched, filtered, unique, output, and newly detected totals plus top sources and queries
• ERRORS — any failed feeds

---

Notes & limitations

• googleNewsUrl is the Google News redirect link (not the publisher's direct URL). Decoding to the final URL is intentionally not attempted — it's fragile and changes often. The source field tells you the publisher.
• sourceUrl is the publisher homepage supplied by the RSS feed, not the final article URL.
• Sentiment is a fast, deterministic headline/snippet signal, not a substitute for a language model or human review.
• Monitoring state is isolated by monitor key plus query/filter configuration. Changing those values intentionally starts a new baseline.
• Google News RSS returns up to ~100 items per feed and favors recent coverage — it's not a full historical archive.
• Results are localized by geo + language.
• For very high volume or repeated scheduled runs, enable Apify Proxy to avoid rate limits.
• Public RSS data only — no login, no paywalled content.

---

Related tools

Part of a Google data-tools suite:

• Google Autocomplete Keyword Scraper
• Google Trends Keyword Monitor
• Google Trends Content Calendar Generator
• Keyword Difficulty Checker

---

Changelog

See CHANGELOG.md.