Built with Hyperbrowser

Minimal CLI that scrapes any URL with Hyperbrowser and compresses it into a semantic knowledge shard (.kzip.json). Great for building growth-data archives from live pages (e.g., auto-curated summaries, link graphs, and key points to power sign-ups and content workflows).

• Get your key at https://hyperbrowser.ai

npm i
npm i -D @types/node

Create a .env in this folder:

echo "HYPERBROWSER_API_KEY=YOUR_KEY_HERE" > .env

npx ts-node kzip.ts <url> [--out file.kzip.json]

# Example
npx ts-node kzip.ts https://news.ycombinator.com --out hn.kzip.json

• Scrape: Uses Hyperbrowser SDK scrape.startAndWait(...) to fetch HTML and metadata
• Compress: Calls extract.startAndWait(...) to generate summary, 5 key points, and outbound links
• Save: Writes <hostname>.kzip.json (or --out), printing compression ratio

{
  "url": "https://example.com/",
  "scrapedAt": "2025-01-01T00:00:00.000Z",
  "title": "...",
  "summary": "...",
  "keyPoints": ["..."],
  "outboundLinks": ["https://..."],
  "rawSize": 12345,
  "compressedSize": 1111
}

• Uses only official Hyperbrowser SDK methods (@hyperbrowser/sdk). No mock data.
• Default output name is <hostname>.kzip.json if --out is omitted.

Follow @hyperbrowser for updates.