# News Scraper & Monitor: RSS Feeds & Keyword Alerts (`koivu-data/news-scraper-monitor`) Actor Monitor news by keyword, website or any RSS, Atom or JSON feed. Finds feeds and news sitemaps, merges duplicates across sources and returns only new articles: title, publisher URL, date, summary, language, matched keywords. Optional full text. Respects robots.txt. - **URL**: https://apify.com/koivu-data/news-scraper-monitor.md - **Developed by:** [Koivu Data](https://apify.com/koivu-data) (community) - **Categories:** News, Automation - **Security:** Limited permissions - **Stats:** 2 total users, 1 monthly users, 100.0% runs succeeded, 0 bookmarks - **User rating**: No ratings yet ## Pricing from $0.80 / 1,000 articles This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events. Since this Actor supports Apify Store discounts, the price gets lower the higher subscription plan you have. Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event ## What's an Apify Actor? An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes. In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours, and optionally produces a well-defined JSON output, datasets with results, or files in key-value store. In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server. Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md. ## How to integrate an Actor? If asked about integration, you help developers integrate Actors into their projects. You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready. Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run. For examples already wired to this Actor's own input schema, see the [API](#api) section below. Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`). # README ![News Scraper & Monitor by Koivu Data: news by keyword, website or RSS feed in one schema](https://api.apify.com/v2/key-value-stores/K8VcTIOdr2cTj8fA7/records/news-hero.png?signature=jJf32Im1aGbWUGStgHmo) **News Scraper & Monitor** collects news articles from the sources you choose and keeps watching them for you. Give it **RSS, Atom or JSON feed URLs**, **news sitemaps** or just **websites** (the Actor finds their feed and news sitemap itself), add **keywords** if you want only some topics, and get every article in **one clean schema**: title, the publisher's own URL, source, publication date, summary, language, matched keywords and where it was found. Built for **monitoring**: turn on *Only new articles*, put the Actor on a schedule, and every run returns only the articles published since the previous run. The same story found in a feed, a sitemap and a second feed of the publisher is **merged into one row**. A run that finds nothing new costs $0.0005 plus $0.0001 per source checked. **Try it now:** the prefilled input reads the newest 20 articles from BBC World and NPR News in a few seconds, for about $0.02. - **Any feed**: RSS 2.0, RSS 1.0 (RDF), Atom and JSON Feed, plus Google-News-format **news sitemaps** - **Websites**: paste `techcrunch.com` and the Actor reads the site's news sitemap and main feed - **Keyword alerts** with whole-word, phrase and prefix matching (`nokia*` also finds *Nokian*, handy for Finnish and other inflected languages), exclude words, languages and *published since* - **Only new articles** across runs, with a separate memory for each saved task. Duplicates are merged across sources - **Optional full article text** where the site's robots.txt allows it. Paywalled articles are skipped, never bypassed - **Polite and transparent**: an honest user agent, robots.txt checked for every URL, no proxies, no browser. Sources that refuse access are reported, not worked around ### What data can you get from news feeds? ![Example output in table view: published date, title, source, language and matched keywords](https://api.apify.com/v2/key-value-stores/K8VcTIOdr2cTj8fA7/records/news-output.png?signature=16zNTkoSiP3om9QINa1Ut) | Field | Example | Notes | |---|---|---| | `title` | `OpenAI will start watermarking ChatGPT's text in the EU` | Headline as the publisher wrote it | | `url` | `https://techcrunch.com/2026/10/05/openai-will-...` | The publisher's own article URL, with tracking parameters (utm\_\*, fbclid ...) removed | | `source`, `sourceDomain` | `TechCrunch`, `techcrunch.com` | Feed or publication name, and the article's domain | | `publishedAt` | `2026-10-05T20:36:48Z` | UTC. From the feed or sitemap (GDELT: when GDELT first saw the article) | | `summary` | `OpenAI will watermark ChatGPT and Codex text in the EU ...` | The publisher's own teaser from the feed, plain text, at most 500 characters. Sitemaps carry no summary | | `language` | `en` | ISO 639-1, from the feed or sitemap, detected from the text if the feed has none | | `author`, `categories`, `imageUrl` | `Aditya Mehta`, `["AI", "OpenAI"]` | When the feed publishes them | | `matchedKeywords`, `matchedIn` | `["OpenAI", "AI"]`, `title` | Which of your keywords matched, and where: `title`, `summary`, `tags` (or `articleText` for GDELT) | | `sourceType`, `feed`, `foundIn` | `rss`, `https://techcrunch.com/feed/` | `rss`, `atom`, `jsonfeed`, `news-sitemap` or `gdelt`; every feed where the same article was found | | `alsoPublishedBy` | `[{"source": "...", "url": "..."}]` | With *Same story* deduplication: other publishers of a near-identical headline | | `changeType`, `firstSeenAt` | `NEW` | Monitoring only: `INITIAL` or `NEW` | | `fullText`, `fullTextStatus`, `wordCount` | `ok`, `930` | With *Fetch full article text* only. Status `ok`, `free_part`, `paywalled`, `robots_disallowed` ... (see Limitations) | | `articleId`, `input`, `scrapedAt` | `9f60551228058223` | Stable ID from the normalized URL, the input line and the time of the run | The dataset has four ready-made views: **Articles**, **Monitoring: new articles**, **Sources and duplicates** and **Full text**. #### Sources the Actor reads (and the ones it does not) | Source | How | Notes | |---|---|---| | RSS, Atom and JSON feeds | The feed URL you give | Any public feed. Unchanged feeds are not downloaded again in monitoring (HTTP ETag / Last-Modified) | | News sitemaps | A sitemap URL, or found in the website's robots.txt | Lists a publisher's recent articles (often the last 48 hours) with titles, dates and keywords. Some sites list far more articles here than in their RSS feed | | Websites | `techcrunch.com`, `https://www.hs.fi` | The Actor reads robots.txt for a news sitemap and the home page for a feed link, then falls back to common feed addresses | | GDELT DOC 2.0 API (experimental, off by default) | Your keywords | Worldwide keyword search in the GDELT Project's open news index. See Limitations: often busy | | Google News | **Not supported** | news.google.com's robots.txt disallows automated access to its RSS feeds, and this Actor follows robots.txt. A Google News URL in *Feeds* is refused with that reason and costs nothing | | Bing News RSS | **Not supported** | Bing's feed terms allow only personal, non-commercial use | ### How to scrape news by keyword, website or feed 1. Add one or more sources to **Feeds, sitemaps or websites**, one per line: a feed URL (`https://feeds.bbci.co.uk/news/world/rss.xml`), a news sitemap URL or a website (`techcrunch.com`). 2. Optional: add **Keywords** to keep only matching articles, for example your brand, competitors or topics. 3. Set **Max articles per run** (newest first). 4. Click **Start**. 5. Download the results as JSON, CSV, Excel or HTML from the run's **Output** or **Storage** tab, or read them through the API. The key-value store record `OUTPUT` says, per source, what was read, how many articles matched and why a source failed. #### Keyword syntax | Keyword | Matches | Does not match | |---|---|---| | `nokia` | Nokia, Nokia's, NOKIA | Nokian, Nokialle | | `nokia*` | Nokia, Nokian, Nokialle, Nokiasta | Snokia | | `electric vehicles` | electric vehicles, Electric-vehicles | electric vehicle | | `electric vehicle*` | electric vehicle, electric vehicles | | Keywords are case-insensitive and are matched in the title, the summary and the feed's tags. Separate several keywords on one line with commas, or put each on its own line. **Exclude keywords** use the same syntax and drop an article when they match. #### How do I find a site's RSS feed? Many sites link their feeds on an "RSS" page or in the page footer. You can also simply paste the website: the Actor looks for the feed link in the home page and for a news sitemap in robots.txt. If it finds neither, the run summary says so and nothing is charged for that source. ### How to monitor news for new articles 1. Save your input as a **task** and enable **Only new articles since the last run**. 2. Create a **schedule** for the task, for example every hour or every morning. 3. Connect an integration to receive each run's new articles: **Slack**, email, Google Sheets, Zapier, Make, n8n or a webhook. ![How Only new articles monitoring works: a saved task, a schedule, and two real runs](https://api.apify.com/v2/key-value-stores/K8VcTIOdr2cTj8fA7/records/news-monitoring.png?signature=lDVSfIUGgc8MGKXyxjsi) The first run returns the current articles that match your filters (marked `INITIAL`) and remembers them. Turn off *Return current articles on the first monitoring run* to start silently. Later runs return only articles that were not seen before (`NEW`). An article counts as seen if its URL (without tracking parameters, `www.`, AMP suffixes and the like) or a near-identical headline from the same publisher was seen before, so a story that moves from one feed to another is not reported twice. If you add a source to a task later, its current articles come once as `INITIAL`, not as a flood of `NEW` rows. If more new articles exist than *Max articles per run*, the newest are returned and the rest are skipped (the summary counts them). If the run stops at your cost limit, the remaining new articles come in the next run. Changing the keywords, exclude keywords, *Apply keywords to feeds*, languages, GDELT countries, an absolute *Published since* date or the duplicates setting starts a new memory (the first run after the change is again `INITIAL`); adding or removing feeds does not. #### Brand and competitor alerts in Slack Monitor the feeds and websites of the media you care about with your brand, product and competitor names as keywords. Schedule the task hourly and connect Apify's Slack integration: each run posts only the new matching headlines with links. A quiet hour costs $0.0005 plus $0.0001 per source. ### How much does it cost to scrape news? This Actor uses **pay per event** pricing. Apify platform usage is included in these prices. | Event | Free and Bronze tiers | Silver tier | Gold tier and higher | |---|---|---|---| | Actor start | $0.0005 per run | $0.0005 | $0.0005 | | Source checked (a feed, sitemap or website read successfully, or one GDELT search) | $0.0001 per source | $0.0001 | $0.0001 | | Article saved (one dataset row) | $0.001 per article ($1.00 per 1,000) | $0.0009 | $0.0008 | | Full text (only with *Fetch full article text*, only when text was extracted) | $0.001 per article | $0.0009 | $0.0008 | Your discount tier comes from your Apify subscription plan. Examples at the Free and Bronze tier prices: - **The prefilled run** (2 feeds, 20 articles): $0.0005 + 2 × $0.0001 + 20 × $0.001 = about **$0.02**. - **Export 1,000 articles** from 10 feeds and sitemaps: about **$1.00**. - **Monitoring 20 feeds every hour** with about 10 new matching articles a day: 720 runs × ($0.0005 + 20 × $0.0001) = $1.80, plus 300 articles × $0.001 = $0.30, so about **$2.10 per month**. - **A morning digest of 50 sources**: 30 runs × $0.0055 = $0.17 per month plus the articles. Never charged: sources that fail (not found, refused, disallowed by robots.txt, not a feed), duplicates merged into another row, articles removed by your keywords or filters, and full text that could not be extracted (paywalled, disallowed or too short). You can cap any run with **Maximum cost per run** in the run options; the Actor stops cleanly at the cap. #### Is there a free news scraper? RSS feeds themselves are free, and a feed reader is enough for reading news. This Actor is not free: it saves you writing and running the parts around the feeds, which are finding feeds and news sitemaps, keyword matching, deduplication across sources, the "only new" memory, scheduling and integrations. Apify's Free plan includes $5 of usage credit every month, which covers about 4,500 articles with this Actor. ### Limitations We prefer to tell you up front: - **No Google News.** news.google.com's robots.txt disallows automated access to its RSS feeds, and we follow robots.txt. If you need Google News results specifically, this is not the right tool. - **Not a search engine of all news.** The Actor reads the sources you list. Keywords filter those sources; they do not search the whole web. The optional GDELT search is the only worldwide keyword search, and it is unreliable (next point). - **GDELT is experimental and often busy.** GDELT's free API asks for at most one request every 5 seconds. In our tests on 6 October 2026 it refused 22 of about 25 requests as rate-limited (HTTP 429), from a home connection and from Apify's servers alike, even with minutes between requests; the three answers it gave contained no articles, even for *nokia* over three days. The Actor waits once, retries once, never works around the limit, and continues with your feeds. In monitoring, the next successful GDELT search covers the time since the last successful one (up to 3 months). GDELT matches the full article text in English translation: write GDELT keywords in English (names work in any language). GDELT returns no summary, and its date is when GDELT first saw the article. - **Feeds show only recent articles.** A feed typically lists the newest 10–100 items and a news sitemap the last day or two. The Actor cannot fetch older articles than the source publishes; for a continuous record, use monitoring. - **Some sites refuse automated readers.** For example, The Guardian's feeds answered our honest user agent with HTTP 403 in testing. Such sources are reported with the reason and are not charged. The Actor never retries with a disguised browser identity, rotating IPs or proxies. - **robots.txt is followed for every URL,** including redirects, feeds and article pages. If a site's robots.txt cannot be read because of a server error, nothing is fetched from that site in that run (RFC 9309). - **Websites without a feed link or news sitemap** cannot be read; paste the feed URL instead. Yle (yle.fi), for example, publishes feeds such as `https://yle.fi/rss/uutiset/tuoreimmat` but does not link them from its home page. - **Full text** is extracted from the article page (structured data first, then the main text). It works on most news sites, but not on pages that load the text with JavaScript. Subscriber-only articles return `fullTextStatus: paywalled` and no text, even if the page source contains it. When a publisher marks only part of an article as paywalled, you get the part a visitor without a subscription sees (`free_part`). Other statuses: `ok`, `robots_disallowed` (robots.txt disallows the page, or the site's robots.txt could not be read or refused access, in which case nothing is fetched there; on 9 October 2026 this applied to npr.org articles from Apify's cloud), `consent_page`, `bot_check`, `no_text` (too little text, often JavaScript pages) and `http_`. In our tests on 6 October 2026, all Helsingin Sanomat articles were subscriber-only, Ilta-Sanomat articles were `free_part`, and Yle, BBC and TechCrunch were `ok`. - **Language** comes from the feed or sitemap. When a feed does not declare it, the Actor guesses from the words of the title and summary for about a dozen European languages; otherwise it is `null`, and such articles are dropped if you filter by language. - **Deduplication** uses normalized URLs and near-identical headlines. Two different articles with the same headline from the same publisher (for example a recurring "Live updates" title) can be merged; use *Same URL only* to avoid that. ### Input ```json { "feeds": [ "techcrunch.com", "https://www.theverge.com/rss/index.xml", "https://feeds.bbci.co.uk/news/technology/rss.xml" ], "keywords": ["OpenAI", "Nvidia", "electric vehicle*"], "excludeKeywords": ["podcast"], "languages": ["en"], "publishedSince": "24 hours", "maxArticles": 100, "onlyNew": false, "dedupe": "article", "fetchFullText": false } ``` | Input | Description | |---|---| | `feeds` | Feed URLs, news sitemap URLs or websites, one per line. | | `keywords` | Words or phrases to keep (see Keyword syntax). Empty = every article. | | `maxArticles` | Maximum articles saved per run, newest first (default 100). | | `onlyNew` | Monitoring: return only articles that are new since the previous run of the same task. | | `returnExistingOnFirstRun` | Monitoring: whether the first run (and a newly added source) returns the current articles (default on). | | `excludeKeywords` | Drop articles that mention any of these. | | `keywordsFilterFeeds` | On (default): keep only feed articles that match a keyword. Off: keep all, and only tag matches. | | `languages` | ISO 639-1 codes, for example `en`, `fi`, `sv`. | | `publishedSince` | Absolute (`2026-09-01`) or relative (`24 hours`, `7 days`). | | `dedupe` | `article` (default), `story` (also merge near-identical headlines across publishers) or `url`. | | `fetchFullText` | Extract the article text where robots.txt allows; paywalled articles are skipped. | | `searchGdelt`, `countries` | Experimental worldwide keyword search with GDELT, optionally limited to publisher countries (`FI`, `US`). | | `stateKey` | Advanced: name of the monitoring memory slot (defaults to the task ID). | ### Output One article from a real run (the list of categories shortened): ```json { "articleId": "9f60551228058223", "title": "OpenAI will start watermarking ChatGPT's text in the EU", "url": "https://techcrunch.com/2026/10/05/openai-will-start-watermarking-chatgpts-text-in-the-eu/", "source": "TechCrunch", "sourceDomain": "techcrunch.com", "publishedAt": "2026-10-05T20:36:48Z", "summary": "OpenAI will watermark ChatGPT and Codex text in the EU to comply with the AI Act. Editing can make the invisible marks harder to detect, it says.", "language": "en", "country": null, "author": "Aditya Mehta", "categories": ["AI", "Government & Policy", "OpenAI", "EU AI Act"], "imageUrl": null, "matchedKeywords": ["OpenAI", "AI"], "matchedIn": "title", "sourceType": "rss", "feed": "https://techcrunch.com/feed/", "foundIn": ["https://techcrunch.com/feed/"], "alsoPublishedBy": null, "changeType": "INITIAL", "firstSeenAt": "2026-10-06T07:07:07Z", "fullText": null, "fullTextStatus": null, "wordCount": null, "input": "techcrunch.com", "scrapedAt": "2026-10-06T07:07:07Z" } ``` Every run also saves a summary to the key-value store record `OUTPUT`: for each source the type (feed, sitemap, website, GDELT), the URLs it was read from, items in the source, items matching your filters, notes (for example "not modified since the previous run") and a clear reason for every source that failed; plus counts of duplicates merged and of articles removed by each filter. ### Use cases - **PR and brand monitoring:** watch the outlets that matter to you for your company, products and people, and get each new mention once. - **Competitor news:** follow competitors' names across trade media and their own newsroom feeds. - **Investor research:** collect company and sector news from business media feeds into a spreadsheet or database. - **AI agents and RAG pipelines:** a clean, deduplicated stream of headlines, summaries and links (optionally full text where allowed) for summarizing, classification or retrieval. - **Alerts to Slack, email or a webhook:** schedule a task with keywords and connect an integration. - **Media archives and newsletters:** keep a running record of everything a set of publishers releases. ### Use it from Python, JavaScript, n8n or AI agents #### News API in Python ```python from apify_client import ApifyClient client = ApifyClient("") run = client.actor("koivu-data/news-scraper-monitor").call( run_input={"feeds": ["techcrunch.com", "https://www.theverge.com/rss/index.xml"], "keywords": ["OpenAI"], "maxArticles": 50} ) for article in client.dataset(run["defaultDatasetId"]).iterate_items(): print(article["publishedAt"], article["title"], article["url"]) ``` #### JavaScript ```javascript import { ApifyClient } from 'apify-client'; const client = new ApifyClient({ token: '' }); const run = await client.actor('koivu-data/news-scraper-monitor').call({ feeds: ['techcrunch.com', 'https://www.theverge.com/rss/index.xml'], keywords: ['OpenAI'], }); const { items } = await client.dataset(run.defaultDatasetId).listItems(); console.log(items); ``` #### n8n, Make and Zapier Use Apify's integration for n8n, Make or Zapier to start the Actor or a saved task and pass the articles to the next step. For monitoring, schedule the task in Apify and let the integration react to each finished run. #### AI agents (MCP) AI agents connected to the [Apify MCP server](https://mcp.apify.com) can find this Actor by searching for *news monitor* or *RSS feed* and call it as a tool to read or monitor a set of news sources. ### FAQ #### Is it legal to scrape news feeds? Publishers offer RSS feeds and news sitemaps so that software can read them. The Actor reads only those public files and pages, identifies itself honestly, follows each site's robots.txt and stops at paywalls, consent pages and bot checks. By default it stores only headlines, the publisher's own short teaser and links. **Headlines and teasers are still the publishers' content:** use them in line with copyright law and the publishers' terms, for example for monitoring and linking rather than republishing. If you turn on full text, you are responsible for having the right to store and use it. If you are unsure, consult a lawyer. #### Does it read Google News? No. news.google.com's robots.txt disallows automated access to its RSS feeds, and this Actor follows robots.txt. Paste the feeds of the publishers you care about instead, or try the experimental GDELT search for worldwide keyword coverage. #### Can it get articles behind a paywall? No. Paywalled articles still appear with their headline, teaser and link if the publisher's feed lists them, but full text is never extracted from them. #### How fresh is the data? Live. Every run reads the sources at that moment. How quickly an article appears depends on the publisher's feed; most news feeds update within minutes. #### Why did a source fail? The run summary (`OUTPUT`) gives the reason for every source: not found, refused by the site (HTTP 403), disallowed by robots.txt, not a feed, or no feed found on the website. Failed sources are never charged. #### Can I get articles older than what the feed shows? No. Feeds and news sitemaps contain only recent articles. Start monitoring now to build a continuous record. ### Support, maintenance and changelog Built and maintained in Finland by **Koivu Data**, a small data tools studio in the EU. - **Found a bug, a feed that does not parse or a missing field?** Open an issue on the **Issues** tab with your run link, so the problem can be reproduced. - **Maintenance:** every parser is covered by unit tests on saved real feeds, sitemaps and pages, and Apify test-runs the Actor every day. - **Data and privacy:** the Actor reads only public feeds, sitemaps and pages and saves results only to your Apify account. Keyword search results from GDELT come from the [GDELT Project](https://www.gdeltproject.org/). - **Changelog** (full history on the **Changelog** tab): - **0.1** (2026-10-06): first version with feeds, news sitemaps, website discovery, keyword filters, deduplication, monitoring, optional full text and experimental GDELT search. An honest review on the **Reviews** tab helps other users decide and tells us what to improve. # Changelog This Actor's version history is a separate document: https://apify.com/koivu-data/news-scraper-monitor/changelog.md # Actor input Schema ## `feeds` (type: `array`): One per line: an RSS, Atom or JSON feed URL (https://feeds.bbci.co.uk/news/world/rss.xml), a news sitemap URL, or a website (techcrunch.com), for which the Actor finds the feed and the news sitemap itself. With no feeds and no keywords, the Actor reads two example feeds. ## `keywords` (type: `array`): Words or phrases to find, one per line (case-insensitive). nokia matches the whole word; nokia\* also matches inflected forms such as Nokian (useful for Finnish); electric vehicles matches the phrase. Keywords filter the articles of your feeds and websites (title, summary and tags). Leave empty to get every article. ## `maxArticles` (type: `integer`): Maximum number of articles saved per run, newest first. You only pay for articles that are saved. ## `onlyNew` (type: `boolean`): For scheduled runs. The first run returns the current articles (marked INITIAL) and remembers them; later runs return only articles that are new since the previous run (marked NEW), also across sources. The memory is kept per saved task. ## `returnExistingOnFirstRun` (type: `boolean`): On by default. Turn off to start monitoring silently: the first run only remembers the current articles and saves nothing. The same applies to a source you add to a task later. ## `excludeKeywords` (type: `array`): Drop articles whose title, summary or tags contain any of these words or phrases. Same syntax as Keywords. ## `keywordsFilterFeeds` (type: `boolean`): On: keep only articles that mention a keyword in the title, summary or tags. Off: keep every article from your feeds (keywords are then only used for the GDELT search and for the matchedKeywords field). ## `languages` (type: `array`): Keep only articles in these languages: two-letter codes such as en, fi, sv, de. The language comes from the feed, the sitemap or GDELT; articles whose language is unknown are dropped when this is set. ## `publishedSince` (type: `string`): Keep articles published on or after this date. Absolute (2026-09-01) or relative (24 hours, 7 days). The GDELT search covers this range (default: the last 24 hours, or since the previous monitoring run). ## `dedupe` (type: `string`): Same article: one row per article even if it appears in several feeds, sitemaps or GDELT (same address, or the same title from the same publisher). Same story: also merge near-identical headlines from different publishers (for example wire stories); the other publishers are listed in alsoPublishedBy. Same URL only: merge only identical addresses. ## `fetchFullText` (type: `boolean`): Also open each saved article on the publisher's site and extract the text. Only where the site's robots.txt allows it; paywalled articles are skipped (status paywalled) and never bypassed. Charged per article with text. You are responsible for having the right to store and use the text (copyright). ## `searchGdelt` (type: `boolean`): Also search the keywords in the GDELT Project's free index of worldwide online news (about 65 languages, last 3 months). Experimental: in our tests in October 2026 GDELT's free API refused most requests as rate-limited (HTTP 429) and its few answers were empty. The Actor then waits once, retries once and continues with your feeds; it never works around the limit, and a monitoring run catches up when GDELT answers again. GDELT matches the article text (write keywords in English; names work in any language) and returns no summary. ## `countries` (type: `array`): GDELT search only: keep news from publishers in these countries. Two-letter codes (FI, US, GB) or English names. ## `stateKey` (type: `string`): Advanced. Name of the memory slot used by monitoring. Leave empty to use the task ID automatically. ## Actor input object example ```json { "feeds": [ "https://feeds.bbci.co.uk/news/world/rss.xml", "https://feeds.npr.org/1001/rss.xml" ], "keywords": [], "maxArticles": 20, "onlyNew": false, "returnExistingOnFirstRun": true, "keywordsFilterFeeds": true, "dedupe": "article", "fetchFullText": false, "searchGdelt": false } ``` # Actor output Schema ## `results` (type: `string`): No description ## `summary` (type: `string`): No description # API You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup. ## JavaScript example ```javascript import { ApifyClient } from 'apify-client'; // Initialize the ApifyClient with your Apify API token // Replace the '' with your token const client = new ApifyClient({ token: '', }); // Prepare Actor input const input = { "feeds": [ "https://feeds.bbci.co.uk/news/world/rss.xml", "https://feeds.npr.org/1001/rss.xml" ], "keywords": [], "maxArticles": 20 }; // Run the Actor and wait for it to finish const run = await client.actor("koivu-data/news-scraper-monitor").call(input); // Fetch and print Actor results from the run's dataset (if any) console.log('Results from dataset'); console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`); const { items } = await client.dataset(run.defaultDatasetId).listItems(); items.forEach((item) => { console.dir(item); }); // 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs ``` ## Python example ```python from apify_client import ApifyClient # Initialize the ApifyClient with your Apify API token # Replace '' with your token. client = ApifyClient("") # Prepare the Actor input run_input = { "feeds": [ "https://feeds.bbci.co.uk/news/world/rss.xml", "https://feeds.npr.org/1001/rss.xml", ], "keywords": [], "maxArticles": 20, } # Run the Actor and wait for it to finish run = client.actor("koivu-data/news-scraper-monitor").call(run_input=run_input) # Fetch and print Actor results from the run's dataset (if there are any) print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}") for item in client.dataset(run.default_dataset_id).iterate_items(): print(item) # 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start ``` ## CLI example ```bash echo '{ "feeds": [ "https://feeds.bbci.co.uk/news/world/rss.xml", "https://feeds.npr.org/1001/rss.xml" ], "keywords": [], "maxArticles": 20 }' | apify call koivu-data/news-scraper-monitor --silent --output-dataset ``` ## MCP server setup ```json { "mcpServers": { "apify": { "type": "http", "url": "https://mcp.apify.com/?tools=fetch-actor-details,koivu-data/news-scraper-monitor" } } } ``` The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer ` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations). ## OpenAPI specification Download the OpenAPI definition: https://api.apify.com/v2/actors/FAiAddUKvIcZ6L4Tn/builds/GbJmCP2qFkGj2mccD/openapi.json