YouTube Tools
Automatic detection
You can simply mention @Hana and describe your request. If the intent is clear, she will automatically run the relevant YouTube workflow.
What it does
Hana can fetch captions from YouTube videos and use them for summaries and Q&A.
Supported operations/arguments
| Operation | Invocation Keywords | Description | Supported arguments |
|---|---|---|---|
| YouTube captions retrieval | youtube, captions | Retrieve captions for a YouTube video in a chosen language. | video_id, language_code (BCP-47, default en) |
Invocation examples
youtubecaptions
@Hana https://www.youtube.com/watch?v=qAF1NjEVHhY&t=197s
Can you identify key highlights from this video?
The above example fetches captions for the specified YouTube video and uses them to answer your follow-up questions.
More copy-paste examples
Common invocations:
@Hana summarize this YouTube video in 6 bullets: https://www.youtube.com/watch?v=dQw4w9WgXcQ
@Hana extract action items with timestamps from this video: https://www.youtube.com/watch?v=abc123xyz
@Hana what are the top 3 recommendations from this video? https://www.youtube.com/watch?v=abc123xyz
Edge invocations:
@Hana use captions language as es and summarize this video in English: https://www.youtube.com/watch?v=abc123xyz
@Hana compare advice in these two videos and list conflicts: https://www.youtube.com/watch?v=vid111 and https://www.youtube.com/watch?v=vid222
@Hana focus only on content after 12:30 in this video and list risks mentioned: https://www.youtube.com/watch?v=abc123xyz
When to use
Use for summarization and Q&A when a YouTube link has useful spoken content and captions.
Troubleshooting
- No captions found: choose a video with available captions or change language code.
- Weak summary: ask a focused follow-up with topic/time-range constraints.
Permissions/limits
- Accuracy depends on caption quality and language availability.
- Private/restricted videos cannot be processed without accessible captions.
High-signal invocation
@Hana summarize this YouTube video in 5 bullets and include timestamps for each point: https://www.youtube.com/watch?v=abc123xyz
Edge-case invocation
@Hana use captions language as es for this video and return output in English with action items: https://www.youtube.com/watch?v=abc123xyz