Built with Hyperbrowser

A blazingly fast web data extractor powered by Hyperbrowser's scraping infrastructure and OpenAI's latest open-source gpt-oss-20b model. Extract structured data from any website with enterprise-grade reliability and zero API costs for AI inference.

• ✅ Zero AI costs - Local inference with open-source models
• ✅ Enterprise reliability - Hyperbrowser handles CAPTCHAs, proxies, rate limits
• ✅ Lightning fast - gpt-oss-20b delivers lower latency and runs on consumer hardware
• ✅ Fully customizable - Modify extraction schemas for any use case
• ✅ Growth-ready - Built for scale with retry logic and error handling

1. Get an API key at https://hyperbrowser.ai
2. Install vLLM to run gpt-oss-20b locally (works on consumer hardware!)

1. Install dependencies:

npm install

2. Set up environment variables:

cp .env.example .env
# Add your Hyperbrowser API key
echo "HYPERBROWSER_API_KEY=your_api_key_here" > .env

3. Start the gpt-oss-120b model server:

# Install vLLM with gpt-oss support (from official Hugging Face instructions)
uv pip install --pre vllm==0.10.1+gptoss \
    --extra-index-url https://wheels.vllm.ai/gpt-oss/ \
    --extra-index-url https://download.pytorch.org/whl/nightly/cu128 \
    --index-strategy unsafe-best-match

# Start the model server (auto-downloads from Hugging Face)
vllm serve openai/gpt-oss-120b

Run the extractor:

npm run start

The tool will:

1. Scrape Wikipedia's list of largest cities
2. Extract structured data using gpt-oss-120b
3. Save results to cities.json

Perfect for data-driven growth teams who need to:

• 📊 Monitor competitor pricing from e-commerce sites for dynamic pricing strategies
• 📱 Track social media metrics across platforms for content optimization
• 💼 Extract job postings for talent acquisition and market analysis
• ⭐ Scrape product reviews for sentiment analysis and feature insights
• 🎯 Gather market data for business intelligence dashboards
• 📈 Auto-generate LinkedIn carousels from scraped industry stats

{
  "cities": [
    {
      "city": "Tokyo",
      "country": "Japan", 
      "population": 37393128,
      "rank": 1
    },
    {
      "city": "Delhi", 
      "country": "India",
      "population": 32941308,
      "rank": 2
    }
  ]
}

Follow @hyperbrowser for updates.