Extract YouTube video transcripts with metadata and save as Markdown to Obsidian vault. Use this skill when the user requests downloading YouTube transcripts, converting YouTube videos to text, or extracting video subtitles. Does not download video/audio files, only metadata and subtitles.
The 'youtube-transcript' skill enables users to extract YouTube video transcripts along with associated metadata and chapters, saving the results in a Markdown format. It is a useful tool for anyone who needs to convert YouTube videos into text, such as for creating notes or generating summaries. By focusing on extracting textual content and metadata, the skill provides a streamlined way to convert video data into an easily digestible format while leaving out the video and audio files themselves. This makes it a great choice for those who are looking for a simple, efficient tool for obtaining transcripts from YouTube videos without the need to download or store large media files.
The key features of this skill include automatic extraction of video metadata, subtitles (including optional chapters), and the conversion of VTT subtitle files into well-organized Markdown documents. With a YAML frontmatter containing metadata such as video title, upload date, and view count, users can expect a comprehensive and structured output. The tool also supports deduplication to avoid redundant transcript entries, ensuring cleaner output. The flexibility of specifying custom filenames and saving to a pre-configured Obsidian vault directory enhances the skill's adaptability for different workflows and preferences. Whether you're working with educational content, tutorials, or any video requiring detailed text data, this skill provides a powerful solution.
Ideal users include content creators, researchers, transcribers, and students who need an automated method for extracting text data from videos, especially for organizing, annotating, and studying YouTube content. It's particularly suited for users familiar with Markdown, Python scripts, and Obsidian vaults, although no special technical skills are required to use the skill effectively.
You can extract a transcript by running the following command: `python scripts/extract_transcript.py <youtube_url>`. If you want to specify a custom filename, you can use: `python scripts/extract_transcript.py <youtube_url> custom_filename.md`.
The transcript includes YAML frontmatter with metadata such as video title, channel name, YouTube URL, upload date, video duration, description (truncated to 500 characters), tags, view count, and like count.
Yes, you need to have yt-dlp installed, which you can do by running `pip install yt-dlp`.
Yes, you can remove duplicate entries by using the command `python scripts/deduplicate_transcript.py <markdown_file>`, which eliminates transcript prefixes that are common in VTT files.
No, this skill only extracts metadata and subtitles. It does not download video or audio files.
Quick Setup:
.claude/skills/Repository
glebis/claude-skills