Skip to content
UXClaim
User Research

youtube-transcript

Fetch transcripts from YouTube videos and prepare summaries.

What youtube-transcript Does

YouTube Transcript is a Claude Code skill that automatically extracts and processes transcripts from YouTube videos, enabling you to capture spoken content without manual transcription. This tool is essential for designers, researchers, product managers, and content creators who need to quickly reference video content, extract key insights, or generate summaries without watching entire videos. Whether you’re analyzing competitor demos, gathering user research from video testimonials, or documenting design presentations, this skill transforms video content into searchable, analyzable text that integrates seamlessly with your AI-powered workflows.

The skill handles the technical complexity of YouTube’s transcript systems, automatically detecting available captions and converting them into structured text format. It’s particularly valuable when combined with summarization capabilities, allowing you to process hours of video content in minutes. For remote teams and asynchronous workflows, this skill bridges the gap between video-heavy communication and text-based documentation.

How to Install

  1. Access Claude Code environment: Open your Claude Code workspace where you have skill installation capabilities.

  2. Locate the skill source: Navigate to the Tapestry Skills repository.

  3. Copy the skill code: Clone or download the youtube-transcript skill files from the repository.

  4. Install in your environment: Follow your Claude Code installation process. Typical steps include:

    • Creating a new skill directory in your skills folder
    • Copying the youtube-transcript files into that directory
    • Running any setup scripts if included in the skill package
  5. Verify installation: Test the skill by providing a YouTube URL and confirming the transcript is successfully retrieved.

  6. Optional configuration: Check for any configuration files (typically config.json or similar) to customize transcript language preferences or output formatting.

Use Cases

  • Design research analysis: Extract transcripts from user interview videos, usability test recordings, or design critique sessions to create searchable documentation and identify recurring themes without rewatching footage.
  • Competitive analysis: Capture transcripts from competitor product demos, webinars, or announcement videos to quickly understand feature positioning and messaging strategies.
  • Meeting and presentation documentation: Convert recorded design presentations, team standups, or client calls into searchable text archives for team knowledge management and onboarding.
  • Content marketing preparation: Process YouTube video content from your own channel or industry experts to repurpose into blog posts, social media content, or training materials.
  • Accessibility and accessibility compliance: Generate text-based alternatives for video content, improving accessibility for team members with hearing disabilities or preference for reading over video consumption.

How It Works

The YouTube Transcript skill operates by interfacing with YouTube’s caption and transcript systems. When you provide a video URL, the skill first extracts the video ID and attempts to fetch available transcripts through YouTube’s API or built-in transcript mechanisms. YouTube typically maintains auto-generated captions (via speech-to-text) and creator-provided manual transcripts, and the skill intelligently selects the most accurate available option, prioritizing user-created transcripts over automated ones.

Once retrieved, the transcript data is parsed and formatted into clean, structured text that removes timestamps, formatting artifacts, and redundant information. The skill can optionally apply natural language processing to generate summaries, identify key speakers, or highlight important segments. The output is delivered as plain text or structured data that integrates with downstream Claude Code skills for analysis, summarization, or storage.

The skill handles edge cases gracefully, including videos with age-restricted content, private videos (where transcripts aren’t accessible), and videos in languages other than English. It respects YouTube’s terms of service by accessing only publicly available transcript data through legitimate methods, making it compliant for professional and research use.

Frequently asked questions

How do I install the YouTube Transcript skill?
Clone the youtube-transcript folder from the Tapestry Skills repository and place it in your Claude Code skills directory. Most installations require no additional configuration—simply ensure you have internet access for fetching transcripts from YouTube.
What types of YouTube videos work with this skill?
Any public YouTube video with available captions or auto-generated transcripts will work. This includes videos with creator-provided captions, auto-generated captions, and videos in multiple languages. Age-restricted and private videos may not have accessible transcripts.
Can it handle videos without captions?
If a video has no captions and YouTube hasn't generated auto-captions (rare for English content), the skill cannot retrieve a transcript. YouTube automatically generates captions for most videos, but older or very short videos may lack them.
How accurate are YouTube's auto-generated transcripts?
Auto-generated transcripts are typically 85-95% accurate for clear English audio with minimal background noise. Transcripts created by video creators are usually more accurate. The skill returns whichever transcript is available, prioritizing manual ones.
Can I use transcripts for commercial purposes?
Using transcripts for research, documentation, and analysis is generally permitted. However, republishing full transcripts verbatim may violate YouTube's terms of service. Summaries and excerpts for professional analysis are typically acceptable.
Does this skill work with live streams or premieres?
It works with completed live streams and past premieres that have been archived with transcripts. Live transcripts for ongoing streams may have latency issues, and not all live content retains transcripts after completion.
How can I combine this skill with summarization?
Pass the transcript output to Claude's native summarization capabilities or chain it with a dedicated summarization skill. Many users pipe transcript output directly to a summary-generation step, enabling single-command video-to-summary workflows.
What's the rate limit for fetching transcripts?
The skill inherits YouTube's standard rate limits. For typical professional use, you can fetch dozens of transcripts per day without issues. For high-volume scraping, implement delays between requests to avoid throttling.

Glossary

Auto-generated captions
Captions automatically created by YouTube's speech-to-text technology. These are available on most videos but may contain errors, especially with accented speech, technical jargon, or background noise.
Transcript
The complete text representation of spoken content in a video, including dialogue, spoken text, and sometimes descriptions of sounds. Transcripts preserve the sequence and context of video content.
Video ID
The unique identifier assigned to each YouTube video, typically an 11-character alphanumeric string found in the URL (e.g., 'dQw4w9WgXcQ'). The skill uses this to locate and fetch the correct transcript.
Timestamp
Time markers in transcripts indicating when specific text was spoken in the video. The skill typically removes timestamps to provide clean, readable text unless specifically requested.
Speech-to-text
The AI technology that converts spoken audio into written text. YouTube uses this to generate auto-captions when creators haven't provided manual transcripts.

More in User Research

All →
User Research

Ditto Product Research

Run customer research, pricing tests, and product validation using 300K+ AI personas with 92% focus group overlap.

Ask-Ditto
User Research

Reddit Intel

Convert Reddit discussions into bilingual Excel reports for competitor analysis, user insights, and market research.

carrielabs
User Research

Research Team Tutorial

Build a multi-agent AI research system using Claude Code that extracts data from papers, performs analysis, writes articles, and reviews its own work.

Tristan578
User Research

UX Discovery Interviewer

AI-powered agent skill that conducts structured UX research interviews, transforming vague product ideas into actionable discovery insights.