youtube

name	youtube
description	Work with YouTube data via TikHub — fetch video info, downloadable stream URLs, captions/subtitles, comments and replies, channel info, and run general/shorts search. Use when the task targets YouTube. Covers the YouTube Web-V2 API.

YouTube (via TikHub)

Deep coverage of YouTube via the Web-V2 API. Exhaustive endpoints: tikhub-find-endpoint "<goal>" --platform youtube.

Setup gate

[ -z "${TIKHUB_API_KEY:-}" ] && echo "Set TIKHUB_API_KEY first (see tikhub-onboarding)."

Coverage: YouTube-Web-V2. (The older YouTube-Web API is de-scoped — reach it via discovery.)

Key endpoints

Goal	Method + path	Key params
Video info	GET /api/v1/youtube/web_v2/get_video_info	video_id, language_code
Video info by URL	GET /api/v1/youtube/web_v2/get_video_info_v2	video_url
Download streams	GET /api/v1/youtube/web_v2/get_video_streams_v2	video_id | video_url
Captions/subtitles	GET /api/v1/youtube/web_v2/get_video_captions	video_id, language_code, format
Video comments	GET /api/v1/youtube/web_v2/get_video_comments	video_id, sort_by, continuation_token
Comment replies	GET /api/v1/youtube/web_v2/get_video_comment_replies	continuation_token
Channel id (from URL)	GET /api/v1/youtube/web_v2/get_channel_id	channel_url
Channel info	GET /api/v1/youtube/web_v2/get_channel_description	channel_id, continuation_token
General search	GET /api/v1/youtube/web_v2/get_general_search	search_query, upload_time
Shorts search	GET /api/v1/youtube/web_v2/get_shorts_search	search_query, upload_time

Example

curl -s "https://api.tikhub.io/api/v1/youtube/web_v2/get_video_info?video_id=dQw4w9WgXcQ" \
  -H "Authorization: Bearer $TIKHUB_API_KEY"

Pagination

Comments, replies, channel feeds, and the *_v2 search endpoints use continuation_token — pass the token from the previous response; stop when none is returned. Each page is billed — cap it.

Hand off to task skills

• Download videos / captions → social-media-downloader
• Analyze a channel → creator-analytics
• Comment mining → comments-analysis
• Search/listening → social-listening
• Large pulls → bulk-data-export

Red flags

• Use get_video_streams_v2 for downloadable media URLs; get_video_info is metadata only.
• A video_id is the 11-char id; pass full URLs to the *_v2 variants.