When someone creates a link post on Tumblr, the Tumblr crawler fetches the page to generate a rich preview card that appears in the dashboard feed. Tumblr’s crawler has historically been inconsistent with its user agent string, does not execute JavaScript, and reads standard Open Graph meta tags to populate link post previews.
This guide covers how Tumblr’s crawler works, which tags it reads, image handling within the 540px-wide dashboard layout, cache behavior, and common issues – including the historically problematic user agent behavior that caused issues with prerender services.