gpu-cluster-monitor
by sounderliu
Web & Frontend DevelopmentBrief
A high-performance engineering tool for deep web scraping.
Description
What This Skill Does
A web scraping tool that runs inside a Docker container using Crawlee and Playwright to extract content from sites with anti-bot protections. Targets JavaScript-heavy platforms like YouTube and X/Twitter. Outputs structured JSON with status, content type, and raw data.
Running Playwright inside Docker bypasses bot detection that blocks standard HTTP-based scrapers on protected sites.
When to Use It
- Extracting transcripts from YouTube videos
- Scraping posts or threads from X/Twitter
- Getting raw text from JavaScript-rendered pages
- Feeding scraped content directly into LLM pipelines
- Collecting video descriptions in bulk from YouTube
Example Workflow
User asks: Get the transcript from this YouTube video URL
- 1
Validates the YouTube video ID from the provided URL
- 2
Runs the Docker container with the target URL as argument
- 3
Crawlee and Playwright load the page and intercept transcript data
- 4
Parses the JSON output from stdout for status and data fields
JSON object with status SUCCESS, type TRANSCRIPT, videoId, and full transcript text
Requirements
- Docker installed and running on the host machine
Installation
clawhub install sounderliu/gpu-cluster-monitorRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.