For developers

News API.

The live search behind plainradar.com as one REST call. It isn't a lookup in a stored index: when you call it, it reads what the world's press is publishing on your topic right then, keeps the stories that are really about it, and returns each one once, as JSON. The API is available on RapidAPI.

How it differs from a keyword news API

• Live. News sites and blogs are read at the moment of your call. You get the last hours, not what a crawler stored yesterday.
• It understands the topic. Ask in plain language: a topic ("artificial intelligence") or a question ("what is new in space exploration?"). Stories are chosen by what they are about, not by whether they contain your words. There are no boolean queries to write, and no articles that only mention the word in passing.
• A complete view. Every call draws on several kinds of sources at once: a constant watch on news sites worldwide, the leading outlets of each country, and the feeds of publishers of every size, down to specialist blogs, in 45 languages. Results are English unless you ask for another language or for all of them.
• Every story is assessed on its own. Before you get a story, an AI check judges it against your topic and leaves out what doesn't fit. What comes back is relevant, not just matching.
• Each story once. The same story from many outlets is grouped into one result, with the number of sources that cover it. The big stories come first, and you don't page through twenty copies of one wire report.
• Noise filters included. Say what you don't want in plain language, such as "Skip opinion pieces and listicles" or "Only news from Europe".
• A briefing, written for you. Ask for it and you get a few sentences on the main news, each citing the stories it comes from.

Follow topics and events

Use it for a standing topic (an industry, a company, a technology) or for one event as it unfolds (an election, a product launch, a court ruling, a natural disaster). Call it on a schedule and you see new stories as they are published, and how coverage of each grows (sources). Within a search, cursor returns only the stories you haven't had yet, plus updates for the ones you have.

One endpoint

GET /v1/search takes your topic and returns the stories. Send it through RapidAPI with your key:

curl --get "https://plainradar.p.rapidapi.com/v1/search" \
  --data-urlencode "query=artificial intelligence" \
  -H "X-RapidAPI-Host: plainradar.p.rapidapi.com" \
  -H "X-RapidAPI-Key: YOUR_RAPIDAPI_KEY"

Here's a real response for artificial intelligence, trimmed to one story:

{
  "query": "artificial intelligence",
  "pages": [{
    "from": "2026-10-11T07:55:07+00:00",
    "to":   "2026-10-12T07:55:07+00:00",
    "stories": [{
      "id": "5494c9ad12e9",
      "title": "‘More Dangerous Than Drugs’: Bill Gates Warns Of Risks Posed By Artificial Intelligence",
      "link": "https://dailycaller.com/2026/10/11/bill-gates-warns-artificial-super-intelligence-...",
      "published": "2026-10-11T19:42:26+00:00",
      "feed": "dailycaller.com",
      "summary": "Technology icon Bill Gates expressed dire warnings about the consequences of ...",
      "image": "https://images.dailycaller.com/image/...",
      "sources": 9,
      "relevant": "yes"
    }],
    "omitted": 47,
    "updates": [],
    "more": true,
    "cursor": "FETYbegTl8_UXbsu:day"
  }],
  "stats": {"stories": 77, "hidden": 0, "seconds": 5.46},
  "filters": {"language": "English", "country": null,
              "search": "artificial intelligence", "noise": []}
}

That call found 77 stories in about five seconds. Each story has its headline, link, source, publication time, summary, image and how many sources cover it.

Parameters

Good to know

• A call reads the news live, so it usually takes a few seconds. Set your client's timeout accordingly.
• Every call counts as one request of your RapidAPI plan, calls with a cursor too.
• When a call can't run, the answer has an error message and a matching status: 400 for a wrong parameter, 410 for a cursor whose search has expired, 429 for too many searches in a minute, 503 when a search of yours is still running or the service is busy.
• Articles belong to their publishers. The API returns headlines, short summaries and links; the full article is on the publisher's site. The Terms apply.

Building an AI agent?

Agents don't need the REST API. Plainradar has an MCP server with the same search as a tool, so Claude, Cursor and other MCP clients can call it directly. See how to give your AI agent live news with MCP.