Paste any YouTube link and get the parts you need: the video ID, the playlist ID, the start timestamp, the channel handle — plus ready-made embed, short and thumbnail links. It understands watch URLs, youtu.be, Shorts, live, embed, the old /v/ path, the mobile and music hosts, a bare eleven-character ID, and a URL typed without https. It reads the timestamp in all three forms YouTube emits, including the 1h2m3s shape the share button uses. Everything is parsed in the page and nothing is fetched, so it tells you what a URL says rather than whether the video exists.
| Format | ID location |
|---|---|
| youtube.com/watch?v=ID | query parameter v |
| youtu.be/ID | first path segment |
| youtube.com/shorts/ID | second path segment |
| youtube.com/embed/ID | second path segment |
| youtube.com/live/ID | second path segment |
| youtube.com/@handle | channel, no video |
watch?v=, youtu.be/, /shorts/, /live/, /embed/, the retired /v/ and /e/, /playlist?list=, /channel/UC…, /@handle, /c/ and /user/ — on the youtube.com, m., music., gaming. and youtube-nocookie hosts. A bare eleven-character ID and a URL with no scheme both work too.
It should not be. YouTube writes the t parameter three ways — 90, 90s and 1h2m3s — and this reads all three. The last is what the share dialog produces when you tick "start at", which is precisely where simpler parsers drop it.
Because that list was not a real playlist. A value starting RD is a session-generated radio mix and LL is your own Liked Videos; neither opens for anyone else, so returning it as a playlist ID would hand you a link that does not work.
hqdefault at 480×360 is generated for every video and always exists. maxresdefault at 1280×720 only exists if the video was uploaded at 720p or higher — request it for an older or low-resolution upload and you get a grey placeholder rather than an error.
No, and it cannot. Nothing is fetched: the page never contacts YouTube, which is what keeps the tool private and instant. It reports what the URL says, not what the server would answer.
Because it does not set tracking cookies until the viewer presses play, which is the difference between an embed that needs a consent banner and one that does not. It serves the same video.
No. The URL is parsed with the browser's own parser, in the page.
YouTube has accumulated about a dozen URL shapes over twenty years and a parser that handles only `watch?v=` fails on most of what people actually paste. The link is parsed with the browser's own URL parser and then matched against the paths that exist: `/watch` with the id in the query, `youtu.be/ID` with it in the path, `/shorts/ID`, `/live/ID`, `/embed/ID`, the long-retired `/v/ID`, `/playlist`, `/channel/UC…`, `/@handle`, and the legacy `/c/` and `/user/` paths. The mobile, music, gaming and nocookie hosts are all accepted, a bare eleven-character id is accepted because that is the second most common paste, and a URL typed without a scheme is accepted because almost nobody types one. Three details separate this from the five-line version. The timestamp parameter comes in three forms — `t=90`, `t=90s` and `t=1h2m3s` — and the share button's "start at current time" option emits the third, so a parser reading only digits loses the timestamp on exactly the links where it was deliberately set. The `list` parameter is not always a playlist: a value beginning RD is a session-generated radio mix and LL is the viewer's own Liked Videos, neither of which opens for anybody else, so reporting them as a playlist id hands back a link that cannot work. And an extracted id is validated against the eleven-character alphabet before being returned, because a malformed path otherwise yields a confident-looking id that 404s. The derived links include both thumbnail sizes with the distinction stated: `hqdefault` exists for every video, while `maxresdefault` is only generated when the upload was 720p or higher. Embeds are built on the nocookie host. Nothing is fetched — the page never contacts YouTube, so it cannot tell you whether a video exists, only what the URL says.