YouTube keyword research

research_youtube reads live YouTube data. Volumes are Google web search volumes — YouTube's own search volume is not available; say so.

Steps

1. Ideas and competition: research_youtube operation: "keyword_research" with query (the seed), limit (up to 50 ideas) and competition_for (how many top keywords get a competition reading; default 3). 1 keyword lookup credit. Competition is low/medium/high: high means median channel ≥1M subscribers or average views ≥1M; low means both under 50K.
2. What ranks: for the 2–3 best keywords, operation: "search_videos" with query — top results with views, channel subscribers, views per subscriber, age and length (1 credit each). Small channels with high views per subscriber mean the topic is winnable.
3. Titles: operation: "title_ideas" with query and limit — new titles written against those ranking now (1 keyword lookup).
4. country sets the market (default the workspace's).

Tell the user the credit total before running several seeds.

Output

A table: keyword, volume, competition level, median views, median channel size, Shorts share, suggested format and title. Rank by volume against competition; flag ideas where AI Overviews cite videos (check with find_youtube_videos_cited_by_ai).

Next

Save chosen videos as topics and script them: plan_youtube_videos operation: "add_topic" with title and target_keyword, then "write_script" (see youtube-citation-strategy). Spot breakout formats → youtube-outliers.