The Tumblr crawler

When someone creates a link post on Tumblr, the platform fetches the page server-side to build the preview card that appears in the dashboard. The crawler identifies itself as Tumblr/{version}, though the version string has changed over time.

User agent inconsistency

Tumblr’s crawler has a history of unreliable user agent identification. Older versions sometimes sent no user agent or a generic one that didn’t identify as Tumblr. This is a well-documented problem for prerender services that use user agent matching to decide when to serve static HTML – if the service doesn’t recognize Tumblr/{version}, your JS-rendered page gets served as-is, and Tumblr sees no OG tags.

No JavaScript execution

Tumblr’s crawler doesn’t run JavaScript. OG tags must be in the initial HTML response. If you use a prerender service, make sure it recognizes the current Tumblr/{version} user agent.

Response requirements

The crawler expects a 200 OK with text/html. Redirects are followed. Pages behind auth or IP restrictions won’t generate previews.

Checking if Tumblr can reach your page

curl -A "Tumblr/1.0" https://example.com/your-page

If your <meta> tags are in the response, Tumblr’s crawler will see them.