Open Graph tag lookup

OpenGraph.io reads a URL and returns its title, description, image, and type without you scraping the page. When a link lands in a queue, a Beam agent requests these tags, checks them against an approved rule for required fields, and writes a clean preview record into your system or notifies the owner. Links that come back with empty tags, redirect somewhere unexpected, or point to flagged domains get held and passed to a person, so nobody publishes a broken or misleading preview. The agent works only where the connector and account permissions allow it.

Link preview cards

Preview cards pair a headline, summary, and thumbnail so a shared link renders cleanly in chat, email, or a feed. A Beam agent watches for new links, retrieves the card fields, and populates the preview wherever your workflow stores it, applying rules about image size or missing thumbnails before anything ships. Cards that fail those checks, or where the source page looks unrelated to the expected topic, route to a human for a quick look. This keeps automated posts consistent while leaving editorial judgment with the people who own the channel.

Scrape fallback for missing tags

Some pages omit Open Graph tags, so OpenGraph.io can read the underlying HTML and infer a title and description instead. A Beam agent notices when the standard lookup comes back thin, requests the fallback, and records the recovered fields against the link. It applies your rule for what counts as usable content and updates the entry or flags it. Pages that return login walls, timeouts, or content that contradicts the original request go to a person rather than being guessed at, keeping recovered metadata trustworthy and clearly separated from confirmed tags.