Links and ingestion
Ingesting external URLs as entities, with metadata, images, and honest search.
The link shape plus the ingestion pipeline turns URLs into structured entities: title, description, site name, canonical URL, lead image, and indexed text — one call per URL.
ransack links ingest --branch <id> https://example.com/articleIngestion extracts publisher metadata (Open Graph, title tags), downloads the lead image into your assets, and files readable page text into a search_text-role field.
The republication posture
Extracted third-party text is indexed, never displayed: it makes links findable in your site's search without republishing someone else's writing. Your own commentary belongs in your own fields (a subtitle-role note renders on cards and rows). Titles, descriptions, and links out are standard link-sharing metadata.
Images
ransack assets import uploads and publishes in one call, converging on --external-id so re-runs never duplicate. Ingestion can also take --image-url to fetch a specific image. Served images get automatic variants (thumb/card/full) with far-future caching.