video-analyzer

What This Skill Does

The Video Analyzer is a powerful multimedia processing tool designed to bridge the gap between raw video content and actionable insights. By integrating yt-dlp for efficient metadata extraction and whisper-cpp for high-performance local speech-to-text, this skill allows OpenClaw to process content from YouTube, Twitter, and TikTok without needing cloud-based subscription services. It handles the heavy lifting of downloading, transcribing, and intelligently structuring information, turning hours of video into concise, readable summaries.

Installation

To add this capability to your agent, run the following command in your terminal: clawhub install openclaw/skills/skills/minilozio/video-analyzer-skill Ensure your local environment supports the necessary Python dependencies via uv as defined in the skill's environment configuration.

Use Cases

• Research & Learning: Instantly summarize long-form educational videos or lectures to identify core concepts without sitting through the entire runtime.
• Content Creation: Extract specific quotes, timestamps, or key talking points from social media clips to repurpose them for blog posts or articles.
• Media Archiving: Automatically download and backup important video or audio content locally to your desktop for offline access.
• Language Learning: Transcribe foreign language content into text to aid in comprehension and vocabulary building.

Example Prompts

1. "Summarize this YouTube link for me: [URL]. I need the key points and actionable takeaways in a markdown format."
2. "I'm looking for the specific part in this video where the speaker discusses the new API release. Can you find the timestamp for me?"
3. "Download the audio from this TikTok link to my desktop so I can listen to it later."

Tips & Limitations

• Quality Control: Use the --quality max flag for critical accuracy, such as transcribing technical tutorials or legal discussions where every word matters.
• Multilingualism: The skill is natively multilingual. If you are a non-English speaker, your agent will detect the transcript's content and translate the summary into your preferred language.
• Performance: Local Whisper processing is hardware-dependent. Long videos will require more time and system resources. If your machine is low on RAM, stick to the normal quality setting for everyday tasks.