Tubask/Docsv0.5.0

get_transcript

Fetch YouTube captions — plain or timestamped. Search for keywords, slice by time range, or paginate through long videos. Does not consume YouTube Data API quota.

When to use

Find every mention of a keyword (search_in_transcript="latency")
Pull exact wording for a citation with timestamp
Read captions from minute 12 to 18 (start_seconds / end_seconds)
Paginate through a very long transcript (offset / limit)
Get raw captions when summarize_video excerpt isn't enough

For a high-level overview, start with summarize_video. Use get_transcript when precision matters.

Examples

Keyword search

get_transcript(video_ref="https://youtube.com/watch?v=...", search_in_transcript="latency", format="timestamped")

Time slice

get_transcript(video_ref="VIDEO_ID", start_seconds=720, end_seconds=1080, format="timestamped")

Pagination

get_transcript(video_ref="VIDEO_ID", offset=200, limit=200, format="plain")

Language preference

get_transcript(video_ref="VIDEO_ID", languages=["en", "en-US"], format="timestamped")

Parameters

Parameter	Type	Default	Description
video_ref*	string	—	Video URL or ID.
languages	string[] \| null	null	Preferred caption track languages in order.
format	plain \| timestamped	plain	timestamped includes [MM:SS] or ISO offsets per segment.
offset	integer	0	Skip first N caption segments (pagination).
limit	integer	200	Max segments per response.
start_seconds	number \| null	null	Only return captions at or after this time.
end_seconds	number \| null	null	Only return captions before this time.
search_in_transcript	string \| null	null	Filter segments containing this keyword (case-insensitive).

Pagination workflow

First call with offset=0, limit=200
If response indicates more segments, call again with offset=200, then 400, etc.
Combine with search_in_transcript to narrow before paginating

No captions available

Some videos disable captions or have none in your requested language. The tool returns a clear error with next_steps — often suggesting another language or summarize_video from description only.

Quota

Free transcript fetches

Transcript requests do not use YouTube Data API units. Use get_transcript freely for quotes and keyword search; save API quota for search and metadata via youtube_query.