YouTube hides tags from viewers, not from anyone who presses Ctrl+U. Any video's tag set is ten seconds away, and the real craft is what you do after reading it, because the obvious move, copying, is the wrong one.
Here are the two reading methods and the research workflow that uses them properly.
Method 1: the page source
On any watch page: View Page Source, then find "keywords". The video's full tag list sits in a meta tag, comma-separated, exactly as the uploader typed it.
Method 2: the overlays
The research extensions from the toolkit comparison put tags on the watch page itself: install one and every video you study shows its set without the source-view ritual. Worth it if tag-reading becomes part of a regular research habit; overkill for occasional curiosity.
What tag-reading is actually for
Not for copying. For vocabulary: the tags of videos winning your topic show how those creators phrase it, and phrasings are the reusable part.
The workflow: pull up the top five videos for your target topic, read their tag sets, and harvest patterns, the recurring variants, the abbreviations, the phrasings you had not considered. Then feed those observations into your own research: check the promising phrasings against real search data, and build your own focused set from what survives.
Read, learn, verify, build. The tags you ship should never be another video's set; they should be your video's description, informed by the niche's vocabulary.
What you will actually find, honestly
| You will see | It means |
|---|---|
| Sloppy or near-empty tags on huge videos | Their reach never came from the tag box: the recurring lesson |
| Phrasing variants you had not considered | The harvest: vocabulary worth verifying |
| Ancient keyword-stuffed walls | A museum of old advice, not a strategy |
| Branded and channel-name tags | Theirs, and off-limits for you |
The first row is the meta-lesson worth internalizing: read enough winners and the correlation between tag quality and success visibly is not there, which corroborates everything this site says about where tags rank in the hierarchy.
The line not to cross
Reading public metadata is fine. Adopting another channel's branded tags, their name, their series names, to ride their traffic is the misleading-metadata pattern, bad for policy and useless in practice, since association was never how the system worked.
Everything else in the workflow is ordinary competitive research: the same read-the-winners habit as studying their titles and formats, applied to the one field they thought nobody could see.
The one-line takeaway: Ctrl+U and search "keywords" shows any video's tags; extensions put them on the page. Read them for the niche's phrasing vocabulary and the recurring lesson that big videos win elsewhere, then build your own set from verified variants, never from copy-paste.