🏙️ Vancouver Apartment Scraper

Hey! This is a project built with Stagehand that automates apartment hunting in Vancouver.

🏢 What it Does

This script automates searching apartments.com for Vancouver properties, filtering by price and bedroom count, and extracting detailed unit information into a structured CSV. It visits multiple property listings in sequence and extracts data from the "Pricing & Floor Plans" section of each property.

Key features:

• Automated filtering by price, bed type, and property type
• Multi-property scraping (visits up to 5 listings by default)
• Detailed unit data extraction (price, square footage, floor plan, availability)
• CSV export of all collected data

🚀 Setting Up

Get everything ready to run:

npm install
cp .env.example .env
nano .env # Add your Google Generative AI API key

⚙️ Configuration

You can customize your search by modifying these variables at the top of main.ts:

// ======= CONFIGURATION =======
const UNIT_TYPE = "1 Bed";  // Options: "Studio", "1 Bed", "2 Beds", "3+ Beds"
const MIN_PRICE = "2000";   // Minimum price filter
const LOCATION = "mount-pleasant-vancouver-bc"; // Location to search
const MAX_LISTINGS = 5;     // Maximum number of properties to visit
// ============================

🏃‍♂️ Running the Scraper

Start the apartment search:

npm start

💡 Personal Note

I created this tool as part of my journey to bring modern technology to my traditional finance role in Real Estate. This script represents a step toward fully automating property research and analysis, allowing me to focus on higher-value activities while letting code handle the repetitive data collection tasks.

🔄 Run on Browserbase

To run on Browserbase, add your API keys to .env and change env: "LOCAL" to env: "BROWSERBASE" in stagehand.config.ts.

🤖 Use Different AI Models

By default, this uses Google's Gemini model with their free API (1000 requests/month free), which I specifically customized to make this project accessible without cost. You can also modify stagehand.config.ts to use:

1. OpenAI: Update model to "gpt-4o" and use your OPENAI_API_KEY
2. Anthropic: Update model to "claude-3-5-sonnet-latest" and use your ANTHROPIC_API_KEY

🙏 Acknowledgements

This project wouldn't be possible without:

• BrowserBase and their browser automation platform
• Stagehand for AI-driven web automation
• Google's Generative AI APIs for intelligent extraction