What can you build with the News API?#

You can build live news feeds, brand and competitor monitoring, breaking-news alerting, market and media intelligence, AI assistants, and research datasets — all on top of the same search endpoint, /v1/news/everything, and the same filter parameters. You learn the filters once and reuse them across every scenario below; the only thing that changes is which filters you combine and how you receive the result (pull, stream, or export).

Each section names the real parameters and endpoints that power that use case, so you can go from idea to a working request immediately. If you have not made a call yet, start with how to run your first news search and come back.

Build a news feed or aggregator#

Pull articles that match a topic, language, or publisher and render them as your own feed. /v1/news/everything is the core query; narrow it with language.code, source.domain, published_at.start / published_at.end, and the title filter for keyword matching (wrap a phrase in quotes for an exact match). The same query can be returned as a ready-made RSS feed by adding export=rss, so an RSS reader or a CMS can consume it with no JSON parsing on your side:

curl "https://api.apitube.io/v1/news/everything?language.code=en&title=climate&export=rss" \
  -H "X-API-Key: YOUR_API_KEY"

Drop export=rss and you get the same articles as JSON for a custom front end. Export also covers csv, tsv, xml, xlsx, parquet, and jsonl.

Monitor a brand, company or person#

Track every mention of a company, brand, or individual and how the coverage leans. Filter the same search with organization.name, brand.name, or person.name, then layer entity.sentiment.polarity (positive, negative, or neutral) to surface only coverage that is negative about that entity — the signal a PR or comms team watches:

curl "https://api.apitube.io/v1/news/everything?organization.name=Tesla&entity.sentiment.polarity=negative" \
  -H "X-API-Key: YOUR_API_KEY"

For an ongoing watch rather than a one-off query, see how to monitor news about a company or brand. You can also have new matches pushed to you instead of polling — over the SSE stream, WebSocket, or a webhook.

Track breaking news and real-time events#

Catch developing stories the moment they are indexed. The is_breaking=1 filter narrows any search to articles flagged as breaking, and /v1/news/stream (Server-Sent Events) or /v1/news/ws (WebSocket) delivers matching articles as a live push feed instead of a poll. Combine the two and you have a real-time breaking-news alerter:

curl -N "https://api.apitube.io/v1/news/stream?is_breaking=1&language.code=en" \
  -H "X-API-Key: YOUR_API_KEY"

The stream accepts the same filters as the search endpoint, so you can scope it to a topic, country, or company. See how to stream news in real time with SSE for handling the connection.

Run market and media intelligence#

Turn coverage into signals. /v1/news/trends surfaces trending topics and entities over time; /v1/news/count returns how many articles match a query without downloading them, which is ideal for share-of-voice and volume charts. Add entity.sentiment.polarity to read tone toward a specific company, source.bias (left, center, right) to compare how outlets across the spectrum cover a story, and the /v1/companies, /v1/people, and /v1/journalists directories to enrich names with profiles. Together these power market intelligence, media analysis, and financial sentiment dashboards.

Power an AI app or agent#

Feed clean article text to a model. /v1/news/raw returns the raw body and HTML of an article, which is the building block for retrieval-augmented generation (RAG) and summarization. /v1/fact-check checks a claim against the news index. For assistants, APITube ships an MCP server that exposes the News API as tools — search_news and suggest — plus ready-made prompts like monitor_company, topic_sentiment, and breaking_news, so an AI client (such as Claude or Cursor) can query the news directly without you writing HTTP code.

Export datasets for research and pipelines#

Build datasets for analysis or machine learning. Any search becomes a file by adding export=jsonl or export=parquet — both stream cleanly into data pipelines and notebooks — while csv and xlsx suit spreadsheets. Pair this with /v1/news/count to size a dataset before pulling it, and with published_at.start / published_at.end to slice it by date range.

Common Questions#

• Do different use cases use different endpoints?
• Can news be pushed to me instead of polling?
• Can I build with AI tools without writing API code?

Do different use cases use different endpoints?#

Mostly no. /v1/news/everything plus its filters powers feeds, monitoring, intelligence, and dataset exports — you change the filters, not the endpoint. A few jobs have their own endpoint because they return a different shape: /v1/news/trends for trending topics, /v1/news/count for match counts, /v1/news/raw for full article text, and /v1/fact-check for claim checking. The full parameter list lives in the official News API parameter reference.

Can news be pushed to me instead of polling?#

Yes. Alongside pull-based REST requests, APITube offers a Server-Sent Events stream (/v1/news/stream) and a WebSocket feed (/v1/news/ws) that push matching articles as they are indexed, plus webhooks that deliver new matches to a URL you control. Use pull for periodic jobs and streaming or webhooks when you need news the instant it arrives.

Can I build with AI tools without writing API code?#

Yes. The APITube MCP server wraps the News API as tools (search_news, suggest) that an MCP-compatible assistant — such as Claude or Cursor — can call directly. The assistant runs the search for you, so you can prototype a news-aware agent without hand-coding HTTP requests against the endpoints.