DEV Community

George Kioko
George Kioko

Posted on

How to Feed Your AI Agent Fresh News on Any Topic, No News API

How to Feed Your AI Agent Fresh News on Any Topic, No News API

If you are building anything that needs to know what is happening right now, a brand monitor, a market research agent, a competitor tracker, you hit the same wall. The agent is smart but blind to anything after its training cutoff. The paid news APIs are expensive and rate limited, and most of them cover a narrow slice of sources.

There is a simpler path. You give a topic or a brand name and get back the latest news on it as structured data, ready to feed an agent or drop into a report. Here is the workflow.

The job, stated plainly

You want fresh news on a topic, a company, a person, whatever you are tracking, and you want it as clean structured rows, not a scraped mess of HTML. Title, source, link, date, enough to act on or hand to a model. And you want it to keep up, not to be a one time pull you run by hand.

The slow way is opening Google News, scanning headlines, copying links into a spreadsheet. That does not scale and it goes stale the moment you stop. The tool way pulls the latest on any query in one call.

Step 1: point the actor at a topic

I use the Google News Monitor on Apify. You give it a search query, a brand, a keyword, a person, and it returns the recent news matching it.

{
  "query": "your brand or topic",
  "maxResults": 50,
  "language": "en"
}
Enter fullscreen mode Exit fullscreen mode

Track a competitor, a product category, a regulation, anything you need to stay current on.

Step 2: get structured rows back

Each article comes back as a clean row: title, source, link, and publish date. That is the difference between data you can sort, filter, and feed to a model, and a page of HTML you have to clean first.

For an agent the structure is the point. A list of titles with sources and dates is context your model can summarize, classify, or alert on. A raw page is just noise.

Step 3: turn it into a brief or an alert

Once you have the rows, the rest is your pipeline. Summarize the day's coverage into a brief, classify articles by sentiment, or alert when a competitor shows up in the news. A query becomes a feed, and the feed becomes something your team or your agent acts on.

This is the use case that keeps coming up: give an agent a live window into the news so it is not blind to the present. The monitor is the step that opens that window.

Step 4: run it inside Claude or any AI agent (MCP)

The actor is exposed over the Apify MCP server, so an agent can pull fresh news mid conversation with no glue code:

https://mcp.apify.com?tools=george.the.developer/google-news-monitor
Enter fullscreen mode Exit fullscreen mode

Ask your agent "what is the latest news on this company and summarize the three biggest stories" and it runs the actor, reads the rows, and answers. The news becomes one more source your agent can check on demand.

Why no news API matters

The paid news APIs cost real money, cap your calls, and often miss sources. This gives you the breadth of Google News behind one query, as structured data, without a contract or a key to manage. That is the difference between a feature you can wire into an agent today and a vendor bill you have to justify.

The 10 minute version

Point it at a topic, pull the latest, and feed it into a brief, an alert, or an agent. After the first run you have a repeatable live news feed on anything you track, without a paid news API.

You can run the Google News Monitor here: https://apify.com/george.the.developer/google-news-monitor

Source and verification reports: github.com/the-ai-entrepreneur-ai-hub/apify-actor-portfolio.

Top comments (0)