Now that you’ve created your account and tried your first query in the Sandbox (step one: create your account, step two: run your first search), you’re now ready for the next step: shaping and refining your results.

Filters are where NewsAPI.ai becomes truly powerful and where your results stop being generic and start becoming yours. They decide what enters your dataset, what gets ignored, and how clean, focused, or noisy your feed will be. Once you understand how each filter works, you can move from broad results to precise news streams that reflect your exact goals — whether you’re analyzing a company, following a crisis, tracking a region, or preparing data for a model.

In this guide, we’ll walk through every filter available in the Sandbox, explain when and how to use them, highlight common mistakes, and show how combining filters leads to clearer, more intentional insights.

1. Understanding the Filter Bar

When you open Get Articles in the Sandbox, you’ll see a row of filter categories beneath the query field:

• Locations
• Sources
• Categories
• Time of interest
• Language
• Miscellaneous

Each filter adds structure and intention to your results. You can use them individually or combine them to build a highly targeted query.

2. Locations Filter

What this filter does — and when to use it

The Locations filter limits results based on where the event in the article took place. It doesn’t filter by where the publisher is based — only the event location, detected from the article’s dateline and extracted location concepts.

Use the Locations filter when your analysis depends on the place of the event, whether it’s a city, country, region, or continent. Typical scenarios include:

• Country-specific monitoring (e.g., elections in France, economic updates in Brazil)
• Regional conflicts or crisis tracking (e.g., Gaza, Eastern Ukraine, Sahel)
• Local policy and business developments (city-level announcements, local markets, infrastructure)
• Geographically bounded topics (protests, extreme weather, natural disasters, energy supply issues)
• Comparing how the same story unfolds across different places (e.g., coverage of heatwaves across Europe)

Any time where something is happening matters, this is the filter to use.

How to use it

Open the Locations dropdown and:

• start typing the name of a city, region, country, or broader area
• select one or more locations you want to include
• optionally combine several locations in the same query (e.g., France + Germany + Italy)

Once selected, these locations will work together with any other filters you set (sources, time, language, etc.).

Common mistake

A frequent confusion is between:

• Event location → where the news happens
• Source location → where the publisher is based

If you want “articles about events in France,” use Locations. If you want “articles from French media, regardless of where the event happened,” use source location under Sources.

Mixing these up leads to completely different datasets.

3. Sources Filter

What this filter does — and why it matters

The Sources filter determines which publishers and authors are allowed into your results. It shapes the media environment behind your dataset — mainstream outlets, regional publishers, niche blogs, topic-focused media, or individual journalists.

You can filter sources by:

• Name – specific outlets (e.g., BBC, Reuters, Bloomberg)
• Group – predefined thematic groups (e.g., general/Business, general/Science)
• Location – where the publisher is based
• Source rank percentile – based on web traffic popularity
• Authors – include or exclude specific journalists (when available)

The rank percentile slider helps you control the breadth of your dataset:

• Lower percentiles (e.g., 0–10) focus on the most-visited, high-profile sources.
• Higher percentiles (e.g., 90–100) include niche and long-tail publishers.
• Using a broad range such as 0–100 keeps all sources included.

Choosing sources is an editorial decision — it strongly influences which narratives, regions, and perspectives appear in your results.

When to use it

Use the Sources filter when you want to:

• Control the credibility or profile of the outlets you include
• Focus on certain media ecosystems (e.g., business media, tech media, science press)
• See how specific regions report the news by filtering publishers from particular countries
• Compare different publisher groups covering the same story
• Exclude domains or groups you don’t want in your dataset

How to use it

Open the Sources dropdown and choose one or more of the following:

• By name – select individual domains
• By group – filter by source categories
• By location – limit results to publishers based in specific countries or cities
• Source rank percentile – adjust the range of publisher popularity
• Authors – include or exclude specific author URIs (when available)

These options combine with your query and other filters to shape who is reporting the news you see.

Common mistakes

• Confusing source location with event location Source location = where the publisher is based. Event location = where the news actually happened.
• Over-restricting by combining many source filters Filtering by specific outlets and strict percentile ranges and groups can make the query too narrow.

Want to understand source selection on a deeper level?

If you want a conceptual explanation of why source choice matters — with real examples showing how it changes narratives even for the same event — the following guide provides a deeper overview:

“Build Better News Datasets: Why Source Filtering Matters in News APIs.”

This complements the practical steps in this section.

4. Categories Filter

The Categories filter lets you restrict results to specific high-level topics from the NewsAPI.ai taxonomy. The top-level categories in the Sandbox include:

Arts / Business / Computers / Games / Health / Home / Recreation / Science / Shopping / Society / Sports

Each category expands into more detailed subcategories, allowing you to fine-tune your topic focus.

Why and when to use it

Use the Categories filter when you want to keep your dataset topically clean and aligned with a specific domain. It’s especially useful when:

• Monitoring a particular sector e.g., Business → Finance → Banking for financial market insights
• Filtering out irrelevant noise e.g., removing Sports, Games, or Entertainment topics from a policy or risk-focused feed
• Tracking industry-specific trends e.g., Health → Public Health during outbreak monitoring
• Supporting dashboards that need clearly scoped content e.g., Science & Technology coverage for innovation teams
• Ensuring your analytics pipeline receives clean, topic-aligned content e.g., feeding only Society + Politics to a political-risk classifier

Categories are not meant to replace search logic — they reinforce your topic boundaries so you get fewer off-topic articles.

How to use it

Open the Categories dropdown to:

• Browse the full hierarchical category tree
• Select one or multiple categories
• Combine categories if your use case spans multiple fields (e.g., Business + Science for innovation coverage)
• Start broad, then drill into subcategories if needed

If no category is selected, your query will span all topics.

Common mistakes

• Over-filtering: Choosing narrow subcategories and a narrow query can produce no results. Fix: Start with a broader category → refine later.
• Using categories when you should use concepts: Categories group articles by general topic, but they don’t replace concept-level entity tracking (e.g., “Tesla, Inc.” or “inflation”).
• Selecting too many unrelated categories: Combining distant topics (e.g., Sports + Health + Finance) dilutes the usefulness of the filter.

5. Time Filter (Time of Interest)

What this filter does

This filter defines when the articles you want were published. It limits results to a specific time window.

Available options in the Sandbox include:

• Last week
• Last month
• Custom date range

Trial plan reminder

Free accounts can access articles from the last 30 days. If you select dates beyond that window, you’ll simply get no results — the API isn’t broken, it’s just outside your plan’s range.

How to use it

You can either:

• pick a quick shortcut (Last week / Last month), or
• set an exact Start date and End date in the custom range.

Common mistake

Selecting a date range older than 30 days while on the free plan — this always returns an empty dataset.

When to use it

Use this filter when you want to focus your query on a specific point in time, such as:

• Time-sensitive analysis (e.g., last week’s market movements)
• Crisis monitoring (e.g., coverage from the last 72 hours of a breaking event)
• Trend detection (e.g., last month’s sentiment shift)
• Monthly or quarterly report preparation

6. Language Filter

What this filter does

The Language filter limits your results to articles written in specific languages. NewsAPI.ai automatically detects article language across 60+ languages, so you simply choose which ones you want to include.

How to use it

Open the Language dropdown and either:

• pick a language from the list, or
• type into the search bar (e.g., “English”, “French”, “Arabic”).

You can select one or multiple languages.

When to use it

Use the Language filter when you want to:

• Avoid multilingual noise in broad global queries
• Compare regional perspectives (e.g., English vs. Spanish coverage)
• Keep your analysis consistent by limiting all results to a single language
• Focus on a specific market or audience (e.g., German-language news for DACH countries)

Common mistake

Running a global topic search without any language filter — this often produces mixed-language results that are harder to read, compare, or analyze.

7. Miscellaneous Filter

What this filter does

The Miscellaneous filter groups several important controls that refine the type and quality of articles you receive. These options help you manage content type, duplication, event clustering, social engagement, and sentiment.

It includes:

A. Data Type (News / PR / Blogs)

Choose which content types you want to include:

• News → clean editorial reporting
• PR → official press releases
• Blogs → commentary, opinion, analysis and governmental sites

This helps you match the data to your use case (e.g., avoiding PR noise or extracting only official statements).

B. Duplicate Control

There are two separate duplicate filters:

• Is-duplicate filter Controls whether the article itself is a duplicate. Options:
  • No filter
  • Show only articles that are duplicates
  • Show only articles that are not duplicates (originals)
• Has-duplicate filter Controls whether an article has other copies in the system. Options:
  • No filter
  • Show only articles that have duplicates
  • Show only articles that do not have duplicates

Use these when you want either a clean, de-duplicated list of originals or when you specifically want to analyze how widely a story was republished.

C. Event Filtering

This lets you decide whether to show:

• Articles that are part of an event cluster
• Articles that are not part of any event
• All articles

This is ideal when you want structured, event-based coverage (clusters) or want to see standalone articles.

D. Social Media Shares

Set the minimum number of shares an article must have to be included.

Use it to focus on:

• high-engagement coverage
• stories that are gaining traction
• early “trending” signals

E. Sentiment Slider

Set a sentiment range from –1 (negative) to +1 (positive).For example, –0.2 to +1 keeps neutral-to-positive coverage.

Important: Sentiment scores are available for English-language articles, using it will return only English articles.

How to use it

All controls appear in a single panel inside the Miscellaneous dropdown.You can adjust multiple options at once — they all combine with your query and other filters.

Common mistakes

• Setting sentiment ranges too narrowly (e.g., 0.8 to 1 → almost always empty)
• Selecting News + PR + Blogs together (This often adds unnecessary noise unless your use case truly requires all content types)

When to use it

Use the Misc filter when you want to:

• perform sentiment-based monitoring
• track emerging stories or trends
• detect crisis signals (high engagement, negative sentiment)
• filter structured vs unstructured coverage via event clusters
• remove unnecessary PR content
• ensure a clean, focused dataset for dashboards or analysis

8. Filtering Strategy (Short, Analyst-Focused)

A few principles make filtering far more effective:

9. Guided Exercise: Build a Fully Filtered Query (Apple Inc.)

Let’s create a real, precise search using all major filters.

Topic

Apple Inc. (Concept Search)

Step-by-step

1. Search query: Type Apple Inc. and select the company concept.
2. Locations: Choose United States to focus on events and developments taking place there.
3. Time of interest: Select Last month.
4. Language: Choose English.
5. Sources:
   • Filter by source rank: 0–20 (top publishers)
   • Optional: filter by group → Business
6. Categories:Select Business → Companies.
7. Misc:
   • Data type: News
   • Duplicates: Original articles only, no duplicates.
   • Sentiment: from –0.2 to +1
   • Event filtering: All articles (default No filter)
8. Click Execute Search.

What you get

A clean, business-focused feed of top-tier English-language coverage of Apple Inc., over the past month, with duplicates removed and sentiment mildly filtered to exclude only strongly negative outliers.

11. Next Steps

You’ve now learned how to shape your results using the full filter set. From here, you can:

• Explore Get Events to see how NewsAPI.ai groups related articles
• Build topic-specific dashboards
• Start integrating enriched metadata into your workflows

Ready to continue?

Open the Sandbox and experiment with your own topic. Want a deeper walkthrough? Book a demo.