home / skills / jaganpro / sf-skills / util-youtube-analyzer

util-youtube-analyzer skill

safe

/util-youtube-analyzer

This skill analyzes YouTube videos by extracting transcripts locally or from captions to summarize, extract insights, and create notes.

npx playbooks add skill jaganpro/sf-skills --skill util-youtube-analyzer

Review the files below or copy the command above to add this skill to your agents.

Files (5)

SKILL.md

2.3 KB

---
name: util-youtube-analyzer
description: >
  Analyze YouTube videos by extracting and processing transcripts. Triggers when
  user provides a YouTube URL (youtube.com, youtu.be) or asks to analyze, summarize,
  or extract insights from a YouTube video. Supports both existing captions and
  local AI transcription via whisper-cpp. Useful for learning from Dreamforce talks,
  Trail Together sessions, and Salesforce tutorials.
license: MIT
metadata:
  version: "1.0.0"
  author: "Jag Valaiyapathy"
  category: "utility"
---

# YouTube Video Analyzer

Agentic workflow for extracting and analyzing YouTube video content locally.

## Prerequisites

Ensure these are installed (via Homebrew):
- `yt-dlp` — Video/subtitle downloader
- `ffmpeg` — Audio extraction
- `whisper-cpp` — Local transcription (+ model at `~/.local/share/whisper/ggml-base.en.bin`)

## Agentic Workflow

When a user provides a YouTube URL:

### Step 1: Extract Transcript
Run the extraction script from this skill's directory:

```bash
${SKILL_DIR}/scripts/yt-transcript.sh "YOUTUBE_URL"
```

**Output:** Transcript saved to `/tmp/yt-transcript-{video_id}.txt`

### Step 2: Read the Transcript
Use the Read tool to load the transcript:

```
Read /tmp/yt-transcript-{video_id}.txt
```

### Step 3: Analyze Based on User Intent

| User Request | Analysis Pattern |
|--------------|------------------|
| "Summarize this video" | Structured summary with key points, takeaways |
| "What are the main topics?" | Topic extraction with timestamps |
| "Tell me about X" | Search transcript for X, provide context |
| "Create notes" | Formatted notes with section headers |
| "Find quotes about Y" | Extract relevant quotations |

## Script Behavior

The `yt-transcript.sh` script automatically:
1. **Fast path:** Fetches existing YouTube captions (no AI needed, instant)
2. **Fallback:** Downloads audio + transcribes locally via whisper-cpp
3. **Cleanup:** Converts VTT format to clean plain text

## Example Interactions

**User:** "Summarize this Dreamforce talk: https://youtube.com/watch?v=..."
**Claude:** Runs script → Reads transcript → Provides structured summary

**User:** "What did they say about Agentforce in this video?"
**Claude:** Runs script → Reads transcript → Searches for Agentforce mentions → Provides context

Overview

This skill analyzes YouTube videos by extracting and processing their transcripts to produce summaries, topic lists, notes, and targeted quotations. It handles both existing captions and local transcription via whisper-cpp when captions are unavailable. It is optimized for learning from technical talks, tutorials, and conference sessions like Dreamforce or Trail Together. The output is practical, timestamped, and tailored to the user's intent.

How this skill works

When given a YouTube URL or an explicit request to analyze a video, the skill runs a transcript extraction script that first attempts to download YouTube captions and falls back to extracting audio and transcribing locally with whisper-cpp. The resulting plain-text transcript is read and searched, then processed according to the user's requested action (summarize, topic extraction, search, notes, quotes). Results include timestamps and concise, structured outputs for rapid consumption.

When to use it

You want a quick, structured summary of a conference talk or tutorial video.
You need topic extraction or a timestamped table of contents from a long video.
You want to search a video for specific terms, quotes, or mentions (e.g., Agentforce).
You prefer local transcription for privacy or when captions are missing.
You need formatted notes or actionable takeaways from educational content.

Best practices

Provide the full YouTube URL (youtube.com or youtu.be) for reliable extraction.
Specify the desired output: summary, topics, notes, or search for precise results.
If privacy matters, rely on the local transcription fallback with whisper-cpp.
Allow the tool to run the script and read the saved transcript file before asking complex follow-ups.
Request timestamps for long videos to quickly jump to relevant sections.

Example use cases

Summarize a 60-minute Dreamforce session into a 5-bullet executive summary with timestamps.
Extract and list the main topics covered in a Salesforce tutorial video with timestamps.
Search a product demo for all mentions of 'Agentforce' and return surrounding context and quotes.
Create formatted study notes from a Trail Together session, broken into named sections.
Transcribe a video without captions locally and produce a searchable plain-text transcript.

FAQ

What happens if the video has no captions?

The skill falls back to downloading the audio and transcribing it locally with whisper-cpp, producing a clean plain-text transcript.

How are timestamps provided?

Timestamps are preserved from the original captions or derived from the transcription process and included with topics, quotes, and summary items for easy navigation.