How a portal comes together.

Three layers — capture, AI synthesis, delivery — bring the news landscape to the portal. Each layer is anchored to the previously validated facts.

— Step 01

Source configuration

For each topic portal we maintain a configured list of sources — weighted by credibility, genre, geography and language. Content is fetched several times an hour, de-duplicated and classified by topic to the IPTC standard.

— Step 02

Semantic clustering

Items reporting on the same event are condensed into a story group based on vector distance, entity overlap and temporal proximity. Every story runs through a documented lifecycle: Emerging → Active → Mature → Archived.

— Step 03

AI synthesis with fact-checking

For active stories the platform produces a factual summary in journalistic style — based solely on statements corroborated across multiple sources, with automated hallucination, plagiarism and hedging checks before release.

Our quality­promise.

[01]

Grounding instead of hallucination.

AI syntheses are generated solely from previously validated, multiply corroborated facts. Statements that are not sufficiently corroborated are discarded. Before release, automated hallucination, plagiarism and hedging checks apply.

[02]

Source and licence attribution.

Every synthesis references the underlying original sources. Imagery is either from licensed sources (in particular Wikimedia Commons under the respective stated licences) or AI-generated and labelled accordingly — licence information always shown below the image.

[03]

Cookieless web analytics.

We use Matomo exclusively in cookieless mode — no tracking cookies, IP addresses are anonymised on capture, and data stays on our own servers in the EU. Details in each portal's privacy policy.

[04]

Advertising transparent, only with consent.

The portals are partly funded through Google AdSense and Google Ad Manager. Personalised advertising is served only after explicit consent via the consent banner (TCF 2.2). Without consent, only contextual advertising is shown — with no tracking cookies.

[05]

Taking copyright seriously.

We respect the TDM reservations (§42h Austrian Copyright Act, §44b German Copyright Act, Art. 4 DSM Directive) of other content providers. Publishers wishing to be included or removed will find the direct route in our TDM reservation.