Source Methodology

Last updated: July 4, 2026

Nuzenio is a source-attributed news discovery platform. It organizes public RSS feeds, approved publisher feeds, Google News RSS results, and eligible video sources so readers can discover stories quickly and continue to the original publisher for complete reporting.

What Nuzenio Publishes

• Article title, publisher name, short RSS summary, category, image when available, publication time, fetched time, and original publisher link.
• Editorial context such as what happened, why it matters, key facts, timeline, what happens next, and source transparency.
• Source-diversity signals that distinguish one attributed publisher from multiple feed signals or multiple publishers.

What Nuzenio Does Not Publish

• Nuzenio does not copy full publisher articles without a license.
• Nuzenio does not create fake demo news cards.
• Nuzenio does not treat AI context as original reporting or independent verification.

Source Selection

Nuzenio prioritizes sources with clear publisher identity, stable RSS feeds, useful timestamps, original links, and a history of publishing news rather than spam, scraping, or misleading content. Submitted sources can be reviewed by an administrator before becoming approved feed sources.

Sources may be disabled if they repeatedly fail, obscure publisher identity, publish spam-like content, break links, or create attribution risk.

Duplicate and Cluster Handling

Nuzenio groups likely duplicate stories by normalized titles, source links, and feed metadata. If the same publisher appears through multiple feed paths, Nuzenio may show multiple feed signals but does not count that as broad multi-publisher coverage.

A story labeled "Single publisher, 2 signals" means Nuzenio saw more than one feed signal, but only one normalized publisher. Multi-publisher comparison appears only when multiple distinct publishers are detected.

AI Context Rules

AI-assisted summaries and explainers use only visible feed content, source metadata, timestamps, category context, and publisher links. The AI layer should not add unsupported facts, quotes, legal claims, medical advice, financial advice, or political conclusions.

Corrections and Review

Readers and publishers can request corrections for attribution, broken links, summaries, duplicate handling, categories, images, labels, or outdated metadata through the contact page or corrections page.