Google News RSS Delta Monitor

Persistent monitoring for official Google News RSS feeds. The Actor creates a baseline on the first successful run and returns only new, removed-from-feed, or updated article metadata on later runs.

This is a monitoring product, not an RSS downloader or article scraper. It stores prior feed state, compares runs, preserves previous/current values, suppresses duplicate changes, and exposes clear run evidence.

What it monitors

• keywords;
• company and organization names;
• official Google News topics;
• locations;
• custom official https://news.google.com/rss... URLs.

The Actor uses official Google News RSS responses only. It does not visit publishers, extract article bodies, use a browser or proxy, call AI, generate sentiment, or use a paid API.

First run: baseline creation

The first successful run for each monitor saves the current visible RSS items as its baseline. Existing articles are intentionally not emitted as new.

The dataset receives one baseline record such as:

{
  "recordType": "baseline",
  "monitorKind": "company",
  "monitorValue": "OpenAI",
  "articlesInBaseline": 82,
  "message": "Baseline created from 82 current articles. Existing articles are not reported as new; later runs emit changes."
}

Keep the same monitorKey and monitoring inputs on scheduled runs. Changing the key creates an independent baseline.

Later runs: deltas only

For every stable RSS article identity, the Actor detects:

• new — present now, absent from the prior visible feed;
• removed — present in the prior visible feed, absent now;
• updated — the same stable item changed one or more tracked fields.

Tracked fields:

• title;
• source;
• publication date;
• direct article URL, only when the RSS item itself exposes one;
• Google News URL;
• GUID;
• normalized description hash.

Updated records contain previousArticle, currentArticle, and a field-level changes array.

Removal meaning: Google News RSS is a rolling window. removed means the item left the currently visible feed. It does not prove that a publisher deleted or retracted the article.

Quick start

{
  "keywords": ["artificial intelligence regulation"],
  "monitorKey": "ai-regulation",
  "language": "en",
  "country": "US"
}

1. Run once and confirm the Baselines view.
2. Save the input as an Apify Task.
3. Schedule the Task with the same monitorKey.
4. Read changes from the Article changes view or connect a webhook to successful runs.

Inputs

Input	Default	Purpose
keywords	[]	Free-form Google News RSS search queries.
companies	[]	Exact quoted company/organization searches.
topics	[]	WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, or HEALTH.
locations	[]	Exact quoted place searches.
customRssUrls	[]	Official Google News RSS URLs only.
language	en	Two- or three-letter feed language.
country	US	Two-letter Google News edition.
monitorKey	default	Persistent state namespace. Keep it stable.
maxMonitors	50	Combined unique-monitor ceiling; advanced maximum 250.
maxItemsPerMonitor	100	Maximum visible items retained from each feed.
maxConcurrency	10	Concurrent direct RSS checks.
requestTimeoutSecs	10	Whole HTTP transaction deadline.
maxRequestRetries	1	Bounded retry count for temporary failures.
maxResponseBytes	2,000,000	Streaming response ceiling per feed.

At least one monitoring input is required. Equivalent duplicate values are normalized, fetched once, and billed once.

Output and status

Dataset views:

• Article changes — new, removed, and updated items;
• Baselines — successful first-run setup;
• Warnings and failures — blocked, failed, malformed, or oversized sources.

The OUTPUT key-value record contains a human-readable run summary with counts, retries, errors, limits, and a dataset link.

Run status is one of:

• complete — all requested monitors were checked;
• empty — checks succeeded but current feeds contain zero items;
• partial — some monitors succeeded and some failed or were blocked;
• blocked — no monitor succeeded because source access was blocked.

A blocked or failed source is never reported as a successful empty feed.

URL provenance and nulls

Every record includes the official RSS sourceUrl. Google News RSS normally exposes a Google News redirect URL rather than a resolved publisher article URL. Therefore:

• googleNewsUrl contains the supplied Google News link when available;
• articleUrl is null unless the RSS item itself supplies a direct non-Google URL;
• missing contracted values remain null and are never invented.

Pricing

Provisional Store model: $0.005 per monitor-checked event.

One event is charged only after one unique monitor was successfully fetched, parsed, persisted, and written to the dataset. Failed, blocked, duplicate, and retry attempts do not create extra charges. Baseline creation and all resulting change rows are included in that monitor check.

Estimated infrastructure is $0.00015–$0.00045 per successful monitor check in normal batches. At the provisional price, estimated publisher margin after Apify's 20% share is approximately 71–77%. Live validation must confirm these estimates before public pricing approval.

Reliability and cost controls

• direct HTTPS only;
• official news.google.com/rss URLs only;
• complete-transaction timeouts, including body streaming;
• byte enforcement while chunks are received, even without honest Content-Length;
• maximum two retries and no infinite fallback paths;
• per-monitor checkpoints instead of full-state rewrites;
• deterministic output independent of HTTP completion order;
• idempotent PPE intent/completion records;
• default 50-monitor and hard 250-monitor boundaries;
• no browser, proxy, AI, paid API, or publisher-page request.

Limitations

• RSS is a current rolling window, not a permanent Google News archive.
• Feed membership and ranking can change without a publisher changing its article.
• A changed feed limit, query, country, language, or custom URL changes what “removed” means. Use a new monitorKey when intentionally changing monitor semantics.
• Google can change its RSS format or availability.
• No full article text, publisher metadata enrichment, resolved redirect URL, sentiment, or summary is provided.

Responsible use

The Actor collects only metadata returned by public Google News RSS feeds. Users remain responsible for complying with applicable laws, Google terms, publisher rights, and their own data-retention obligations.