Sources enter through RSS, API, or feed adapters where available.
Sources
Where Fracture gets its story signal.
Fracture tracks configured news sources, preserves outlet metadata, and compares coverage only when enough source context exists to make the analysis useful.
Outlet metadata includes source identity, region, reliability, and broad lean priors used for source-map placement.
Articles are grouped into story clusters before Fracture calculates divergence.
Fracture does not scrape blindly.
Each tracked outlet is represented as a configured source with a known slug, name, URL, tier, and metadata profile. New sources should be added through the ingestion layer rather than pasted into the UI.
Feeds and APIs are normalized.
Different outlets publish different payloads. Fracture normalizes article title, URL, summary, publish time, image, and source identity before analysis.
Source maps use outlet priors and article signals.
The source map is not a popularity ranking. It shows the source spread for a story using outlet metadata and the article-level framing signals available in that cluster.
Transparency improves as source coverage grows.
As more outlets and licensed feeds are added, the Sources page can evolve into a fully live source directory backed by the same API as the ingestion system.
Tracked sources
Loading the current source directory.
Fracture keeps the product focused on source comparison, framing distance, and readable context instead of performative certainty.