NEWS PIPELINE
An automated content factory. It ingests a link, a text seed or a monitored news feed, runs every story through a 15-step AI pipeline — translation, live web fact-checking, enrichment, deduplication, rewrite, cover generation and an LLM quality judge — and serves publication-ready articles in any number of languages through a secure REST API.
SIGN IN»The platform continuously monitors external news sources — RSS and Google News feeds, direct links, raw text seeds. Every incoming item is fingerprinted with vector embeddings, so duplicates and near-duplicates are filtered out before they enter the pipeline. Untrusted media are excluded by a configurable domain policy.
A fifteen-step AI conveyor translates each story, verifies its claims against live web sources, enriches short items with research, rewrites the text for originality and generates a matching cover image. An LLM judge scores every language version before the article is categorized and linked to related stories.
Finished articles — title, summary, HTML body, SEO fields, tags, cover and a two-level category tree — are delivered through a token-protected REST API. Downstream sites pull ready-to-publish content the moment it clears quality control, with webhooks announcing every new story.