A pipeline for news, not a CMS.
Centralized RSS ingestion, AI article rewriting, image remastering, and JSON delivery to your consumer sites — one feed in, clean publish-ready content out.
From feed to ready, in five hops.
RSS poll
Each source is polled on its own schedule. New items land in raw_articles.
Extract
cURL with a scrape.do fallback, then a Readability port pulls the article body.
Policy + image strategy
Routing decisions and image plan are made before the rewrite spends tokens.
Competitor scrub
Rewritten with internal-link caps applied and competitor mentions removed.
Remaster
rembg + gpt-image-1, or just gpt-image-1, depending on the source asset.
JSON, bearer-auth, rate-limited.
-
GET /v1/healthliveness check (auth required) -
GET /v1/articles?since=&until=&category=&limit=&cursor=paginated list of ready articles -
GET /v1/articles/{id}single article by id
Authenticate with Authorization: Bearer <api_key>. Scopes
are read_articles and read_drafts. Each key is
rate-limited to 60 requests/minute; remaining quota is returned in
X-RateLimit-Remaining.
curl -H "Authorization: Bearer $API_KEY" \
https://news.endpointr.com/v1/articles?limit=10