Skimmr News - https://newssite.skimmr.ai
Machine-readable retrieval guide for the public site structure.
Updated: 2026-05-22T21:18:32Z

## Site Identity / Description
Skimmr NewsSite is a static news retrieval site designed for AI crawlers, agents, and other machine clients.
It organizes article resources around IPTC topic taxonomies, location taxonomies, sitemap discovery, and parallel retrieval formats.
This file describes the public retrieval paths and semantic structure of the site.

## Status & Coverage
This is a beta. Current coverage focuses on significant parts of Germany and Switzerland, with additional coverage in Norway. Geographic coverage will expand over time.

For questions, contact info@skimmr.ai.

## Data Quality & Provenance
Sources are selected through human-in-the-loop (HITL) curation. Only primary documents are included - verified non-editorial content. Non-informational page types (shopping, careers, comments sections, and similar) are manually excluded to reduce noise.

All content is retrieved from publicly accessible sources only. No content behind logins, paywalls, or access restrictions is included. Sources span a range of institutional types, including companies, governmental bodies, private institutions, and research organizations.

Content is then additionally filtered rule-based and with the help of AI.

The aim is to provide only truly new content.

## Access
Public retrieval uses normal site URLs.
Authentication, crawler policy, and access controls are the same as for the rest of the current site.
This file describes public retrieval only. It is not an API documentation, webhook documentation, or an internal integration guide. Skimmr offers the news content also as JSON-LD by webhooks (https://skimmr.ai/).

## Formats
Each article resource is available through parallel public URLs. Except for the Teaser file they all offer the same content.
/articles/{slug}/
/articles/{slug}/index.md
/articles/{slug}/index.json
/articles/{slug}/teaser.json

The HTML page is the canonical browseable representation of the article resource.
The Markdown file provides clean, low-clutter article text.
The JSON-LD file provides machine-readable structured metadata.
The teaser file provides the limited free preview representation for discovery.
These are the real public retrieval paths for article resources.

## Navigation
/sitemap.xml
/sitemap-structural.xml
/topics/
/locations/
/llms/topics/{slug}.txt
/llms/locations/{slug}.txt

The sitemap index at /sitemap.xml is the primary sitemap entry point for machine discovery.
Topic and location indexes provide additional semantic discovery paths.

## Taxonomy / Semantic Structure
The site is organized around multiple machine-meaningful axes.

Topics are organized under /topics/ and reflect the implemented IPTC MediaTopics taxonomy hierarchy.
Locations are organized under /locations/ and reflect the implemented geographic hierarchy.
Topic pages and location pages provide structured discovery paths to related article resources.
Topic and location term pages also expose machine-readable llms files under /llms/topics/{slug}.txt and /llms/locations/{slug}.txt.
Topic llms files expose topic identity metadata, including IPTC identifiers and sameAs links.
Location llms files expose location identity, hierarchy, and recent teaser links where available.

This structure is intended to support efficient machine retrieval by:
- grouping semantically related articles under shared topic paths
- grouping geographically related articles under shared location paths
- exposing consistent public format paths for each article resource
- linking discovery layers through sitemap and llms files

The sitemap index at /sitemap.xml links to the structural sitemap and generated topic/location sitemap files.

## Teaser / Free Preview
/articles/{slug}/teaser.json

The teaser file is always free.
It provides summary and metadata for relevance assessment before retrieving the full article resource.
The teaser does not include articleBody.
Per-topic and per-location llms files list recent teaser URLs so machine clients can discover article resources efficiently.