blink-new Blink Scrape
Fetch web pages and extract structured information using blink scrape. No project key needed.
Get clean text from a page
blink scrape "https://example.com" --text
Extract specific data with AI
blink scrape "https://news.ycombinator.com" --extract "top 10 story titles and their URLs"
Extract product prices
blink scrape "https://shop.example.com/products" --extract "all product names and prices"
Extract contact information
blink scrape "https://company.example.com/contact" --extract "email addresses and phone numbers"
Get raw HTML/JSON response
blink scrape "https://api.example.com/data"
Machine-readable JSON output
blink scrape "https://example.com" --extract "main headline" --json
Command signature
blink scrape <url> [--text] [--extract <instructions>] [--json]
url— the page to scrape (required)--text— return clean text (strip HTML tags)--extract <instructions>— use AI to extract specific data--json— return JSON output
Tips
- For structured data (tables, lists, prices): use
--extract - For article text: use
--text - For APIs that return JSON: no flags needed
- AI extraction uses
google/gemini-3-flash(fast + accurate for extraction tasks)