Ingestion registry

Three ways in.

Each adapter discovers content differently, but all emit the same normalized article record. Direct publisher URLs win during deduplication; XiaoHu remains a useful broad discovery layer.

01

Anthropic

EN · sitemap+html

Official Anthropic sitemap discovers news, research, engineering, and feature pages; article metadata supplies title and date.

Inspect endpoint

healthy

30

items emitted

02

OpenAI News

EN · rss

Official OpenAI RSS feed. Each item already points to the canonical publisher URL.

Inspect endpoint

healthy

30

items emitted

03

XiaoHu English

EN · rss+html

RSS discovers broad AI entries; the English article page is parsed again to recover the Read the original URL.

Inspect endpoint

healthy

25

items emitted