Videos

The video command resolves one or more videos to their full metadata in one shot. It bootstraps a session from the watch page, then combines the HTML, the /player response (description, dates, category, tags, formats, captions), and the /next response (chapters, related videos, comment token) into a single record.

ytb video dQw4w9WgXcQ

On a terminal this prints a table. Piped, it emits JSONL. Override with -o (table, json, jsonl, csv, tsv, url, id, raw). See Output formats for the full list, and the CLI reference for every flag.

What it returns

The resolved record carries the fields you see in the browser: title, channel name and id, view count, like count, publish and upload dates, duration, category, tags, the full description, thumbnails, and the live/short markers. The --player-derived sections (formats and caption tracks) and the /next-derived sections (chapters, related videos) are attached too, and the sub-flags below pull each of those out as its own row stream.

Narrow the table columns with --fields (lowercase column names, not JSON keys):

ytb video dQw4w9WgXcQ --fields title,channel,views -o table

Many ids, and stdin

video takes more than one id or URL. Mix bare ids and full watch URLs freely:

ytb video dQw4w9WgXcQ "https://www.youtube.com/watch?v=9bZkp7q19f0"

Pass - to read ids or URLs from stdin, one per line. This is how the other commands chain into video:

ytb search "go programming" -o url | ytb video -

-j/--workers sets how many detail fetches run at once when you pass many ids.

What one read already carries

Chapters, the description's links and hashtags, the thumbnail URLs and the caption track list all come back on a plain read and need no flag. -o json has them nested on the record.

The flags below are the parts that cost another request, so each one is a decision rather than a default:

ytb video dQw4w9WgXcQ --captions      # caption tracks whose URLs fetch, one request
ytb video dQw4w9WgXcQ --formats       # the stream list, one request
ytb video dQw4w9WgXcQ --thumbnails    # a HEAD per constructed rendition

--captions is a narrower thing than it sounds. The plain read already lists the tracks; what it lists are the watch page's URLs, and those answer HTTP 200 with an empty body. The flag re-reads the list through the mobile player, whose URLs actually fetch, so the count is the same and the base_url is the part that changed.

--formats stays off by default because it costs a mobile player request per video, and a crawl of a thousand videos should not make a thousand extra calls nobody asked for.

--transcript fetches the transcript text and attaches it to the record, and --lang picks the language:

ytb video dQw4w9WgXcQ --transcript --lang es

--microdata adds what the page's own schema.org markup says, beside what ytb parsed, which is how you check one against the other.

Faster

--no-player never calls the mobile player, whatever else was asked, and resolves from the HTML bootstrap only. It is faster, at the cost of the formats, the caption list, and some player-only fields:

ytb video dQw4w9WgXcQ --no-player

Sibling commands

Several commands print one part of a video on its own, which is handy when that part is all you want:

The related-videos shelf is the one part of the watch page that does not land on the video record, because it is twenty other videos rather than a fact about this one. related prints it, chapters prints the chapter markers, and captions prints the caption track list:

ytb related dQw4w9WgXcQ
ytb chapters GlYgs6v2YfU
ytb captions dQw4w9WgXcQ

chapters names where each chapter came from. markers means YouTube served a chapter list of its own, so the site draws them on the scrubber; description means somebody typed 1:23 Verse 2 and the list is only as good as their typing.

formats lists the muxed and adaptive formats from /player streamingData, deduped by itag. It lists metadata only and does not resolve playable URLs. Filter by track type with --audio or --video, or show only progressive streams with --muxed:

ytb formats dQw4w9WgXcQ --muxed
ytb formats dQw4w9WgXcQ --audio

Persisting

These commands stream and keep nothing. To collect a video and what it links to, run ytb crawl <id>, which writes the records and the claims into the local store. See The store.