A local AI tool for downloading and analyzing video content from Bilibili, Douyin, and Toutiao using frame extraction and local vision models.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install video-analyzer-cn
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install video-analyzer-cn using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Video Analyzer?
The Video Analyzer is a sophisticated tool designed for the Openclaw Skills ecosystem that enables automated processing and summarization of online video content. By combining local vision models with powerful media processing utilities, it allows developers and researchers to gain deep insights into video data without relying on external cloud services.
This skill automates the entire pipeline from URL recognition to final synthesis. It handles the nuances of different platforms, ensuring that whether you are analyzing a short clip from Douyin or a long-form tutorial on Bilibili, the resulting summary is accurate and context-aware. Using Openclaw Skills for this task provides a private, cost-effective alternative to traditional video analysis methods.
Video Analyzer Use Cases
- Summarizing educational content or tutorials from Bilibili to save time during research.
- Extracting visual descriptions and key trends from Douyin short-form videos for marketing analysis.
- Analyzing news or informational videos on Toutiao to generate quick textual summaries.
- Creating archival descriptions for video libraries using local AI to maintain data privacy.
How Video Analyzer Works
- The user sends a video URL which is identified by the skill to determine the source platform.
- The system downloads the video using platform-specific scripts, utilizing tools like yt-dlp or browser-based extraction for restricted links.
- FFmpeg is invoked to extract frames at dynamic intervals based on the total video length (e.g., every 3, 5, or 10 seconds).
- Each extracted frame is analyzed locally using the minicpm-v model via Ollama to generate descriptive metadata.
- The skill aggregates the frame-by-frame descriptions into a coherent and comprehensive summary of the video content.
Video Analyzer Setup
To get started with this skill in your Openclaw Skills environment, ensure the following dependencies are installed:
# Install yt-dlp for video downloads
pip install yt-dlp
# Ensure ffmpeg is installed on your system
# Windows: Add ffmpeg to your Path
$env:Path = "C:\path\to\ffmpeg;$env:Path"
# Pull the vision model via Ollama
ollama pull minicpm-v
# Ensure Python and Chrome are available for Douyin extraction
Video Analyzer Data Schema & Taxonomy
The skill manages its operations within a dedicated workspace, organizing data as follows:
| File Type | Storage Pattern | Purpose |
|---|---|---|
| Video Media | *.mp4 |
Temporary storage for the downloaded source video. |
| Image Frames | frame_*.jpg |
Scaled-down JPEG frames (640px) used for AI inference. |
| Analysis Logic | analyze_*.py |
Generated Python scripts for frame-by-frame identification. |
| Temp Files | ~/.openclaw/workspace/tmp/ |
Central directory for all intermediate processing artifacts. |
Video Analyzer Advanced Features
- Intelligent Frame Rate Control: Automatically adjusts the sampling rate (FPS) based on video duration to balance speed and accuracy.
- Local AI Privacy: Utilizes minicpm-v running on Ollama, ensuring that video data never leaves your local machine.
- Browser-Integrated Extraction: Supports Chrome-based URL extraction for Douyin to bypass complex anti-scraping and timestamp-sensitive links.
- Automated Resource Optimization: Scales frame resolution and JPEG quality to ensure high-speed processing even on limited hardware.
SKILL.md
Loading
Related Openclaw Skills
Resilient Claude Agent
A robust execution layer that decouples Claude Code from the orchestrator using tmux for persistent, crash-resistant coding tasks.
TDD E2E PR Workflow
Automate the end-to-end development cycle from issue selection to pull request creation with mandatory visual verification.
Extract Design System
Reverse-engineer public website design primitives into localized design tokens and CSS variables automatically.
Ark Seedance Video Generation
An end-to-end video generation skill for AI agents using Volcengine Ark Seedance models for professional video synthesis.
Polymarket Micro Session Edge Trader
An automated trading agent designed to exploit predictable mean-reversion patterns in Polymarket 5-minute crypto markets during global session opens.
Compliance Guardian — FinCEN Real Estate Reporting
A specialized compliance tool for real estate professionals to navigate FinCEN reporting requirements for non-financed property transfers.
Featured*
