skills.rest
mathiaschu
mathiaschuCOMMUNITY
Explore all skills by this author →
𝕏@mathiaschuGitHub

watch

Download videos and convert them into timestamped frames and transcripts.

92|10|Updated May 29, 2026
Search Tags:#video#captions#transcription#yt-dlp#whisper#ffmpeg#browser-cookies
One-click install
npx skills add https://github.com/mathiaschu/watch --skill watch-mathiaschu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: watch
Source: https://github.com/mathiaschu/watch/tree/main
Command: npx skills add https://github.com/mathiaschu/watch --skill watch-mathiaschu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires yt-dlp, ffmpeg, mlx-whisper, openai-whisper, and includes scripts (resource) components.

What problem does it solve?

It removes the friction of understanding videos by turning a URL or local file into grounded, timestamped evidence that an AI can inspect and answer questions from.

Core Features & Use Cases

  • Downloads videos from a wide range of supported sites, including YouTube, Instagram, X, Vimeo, TikTok, Loom, and many other yt-dlp sources.
  • Extracts frames with ffmpeg at an auto-scaled rate so short clips are dense and long videos stay within a manageable frame budget.
  • Pulls captions when available and falls back to local Whisper transcription on the user's machine when captions are missing.
  • Supports focused time ranges for questions about a specific moment, plus browser-cookie access for login-gated or private sources.
  • Example use case: ask what happens at a specific timestamp, summarize a long clip, or inspect on-screen text and spoken dialogue together.

Quick Start

Ask Claude to watch a video URL or local file and answer your question about what happens in it.

Frequently Asked Questions about watch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I transcribe a video and extract frames to understand what happens in it?▼

You can transcribe and extract frames by downloading the video with yt-dlp, pulling available captions, or falling back to local Whisper transcription, then using ffmpeg to generate timestamped frames for AI inspection.

Can I analyze videos from YouTube, TikTok, and Instagram without an API key?▼

Yes, you can download and analyze YouTube, TikTok, Instagram, X, Vimeo, and Loom videos without an API key by using yt-dlp as the extraction engine for your video URL.

Do I need to install yt-dlp and ffmpeg to download and timestamp video frames?▼

Yes, yt-dlp and ffmpeg are required dependencies for downloading the source video and extracting auto-scaled timestamped frames, while mlx-whisper or openai-whisper handles missing captions locally.

What is the best way to summarize a specific time range of a long video?▼

The best way to summarize a specific video segment is to use focused time range extraction, which downloads only the targeted clip, transcribes it with Whisper if needed, and generates dense timestamped frames for that exact moment.

How do I access login-gated or private videos for transcription and frame extraction?▼

You can access login-gated or private video sources for transcription by utilizing optional browser cookies, which authenticate your session with yt-dlp to download the protected content without an external API key.

Related Skills

Explore More Content & Communication Skills→
hoanghd218hoanghd218
3

mkt-video-url-to-transcript

Convert video URLs into MP4 files and Whisper-generated TXT, SRT, and JSON transcripts.

Community
Intermediate
Wanli-LeeWanli-Lee
1

video-watcher

Fetch cleaned subtitle transcripts from YouTube or Bilibili videos.

Community
Basic
joegoldinjoegoldin
2

watching-videos

Download videos and analyze frames and transcripts with yt-dlp and ffmpeg.

Community
Advanced
xingBaGanxingBaGan

video-understand

Analyze video content and transcribe audio locally with ffmpeg and Whisper.

Community
Intermediate