What the site collects

• The site fetches publicly available articles from selected RSS feeds and source pages.
• For each article, it stores metadata such as title, link, source, language, publication date, summary, and image URL when available.
• Original publishers remain the source of record. Article links point readers back to the publishers' websites.

How AI is used

• AI is used to translate non-English titles or summaries, group related articles into topic clusters, and identify broad areas of agreement or disagreement.
• The site also uses AI-assisted analysis to surface narrative divergence, framing differences, missing context, source signals, and topic evolution over time.
• AI-generated analysis can be incomplete or wrong. Readers should verify important details against the linked original reporting.

How much human input there is

• Human input is used to choose the overall source list, maintain the website, design the interface, define categories, and shape the prompts and ranking logic.
• The live topic clustering and article-level analysis are generated automatically from the available source material.
• Unless a page says otherwise, individual topic summaries and source comparisons are not manually edited before publication.

What the site is not

• NewsUnbound does not claim to determine the final truth of a story.
• It does not endorse every article, outlet, claim, or viewpoint it links to.
• It is intended for media literacy, context, and comparison, not legal, financial, medical, or security advice.