A single TikTok video by its URL, with caption, cover, date and counters.

One public TikTok video by its URL, read logged-out from the video page itself. Takes the FULL video address -- https://www.tiktok.com/@handle/video/1234567890123456789 -- and not an id on its own: the document lives under the handle, so an id without it cannot be addressed. items carries exactly one entry: the video id, its page URL, the caption in text, the cover image in thumbnail_url, the publication date parsed from an epoch in seconds, and metrics with likes, comments, shares, views and saves. title is null -- a video has no title, its caption is a body -- and position is null too, because a list of one has no meaningful order. author carries the id, handle, name, bio, avatar and verified flag the video page publishes, and followers, following and posts_count come back null ON PURPOSE: this page gives a follower count with no exact counterpart to check it against, while tiktok/profile reads the exact one; filling both from different sources would make two of our endpoints disagree about the same account. WARNING: THE LARGE COUNTERS ARE ROUNDED BY TIKTOK on this surface -- measured on 2026-10-02 against the committed capture, both of the page stat blocks agree on 2,000,000 likes and 5,500,000 views, so treat them as rounded to three significant figures and not as exact. A video that is restricted by its owner and an id that does not exist both answer 404 and are NOT charged: the page returns the same status code for the two, and only the status message tells them apart. has_more and next_cursor are null: the document is one video, there is no list to page. Costs 1 credit and caches for an hour -- the video itself does not change, and the counters that do are already rounded.