facebook-marketplace-scraper

name	facebook-marketplace-scraper
description	Scrape Facebook Marketplace listings using Playwright and BeautifulSoup. Supports configurable searches by city, query, and maximum price. Requires a one-time login to save authentication state. Includes relevance filtering based on query keywords and a debug mode. Use when you need to extract structured data from Facebook Marketplace for research or personal use.

Facebook Marketplace Scraper Skill

This skill allows you to search and extract listing information from Facebook Marketplace.

Setup (One-time)

Before you can scrape, you need to log in to Facebook and save your authentication state. This generates an auth_state.json file that the scraper uses.

1. Run the login script:

   python3 login_and_save_cookies.py
   

2. A browser window will open. Log into Facebook manually.
3. Once you are on the Facebook homepage, return to your terminal and press Enter.
4. The script will save your login session to auth_state.json.

Important: The .gitignore file has been configured to prevent auth_state.json from being committed to your repository.

Usage

Once authenticated, you can run the scraper with your desired search parameters. The script currently focuses on fetching listings from the initial page load for improved reliability, as dynamic scrolling was found to be less stable.

python3 main.py --city "YOUR_CITY" --query "YOUR_SEARCH_QUERY" [--max_price MAX_PRICE] [--no-headless] [--debug]

Arguments:

• --city (required): The city for the search (e.g., "Richmond", "Vancouver").
• --query (required): The search query (e.g., "iPad Air", "vintage bike").
• --max_price (optional, default: 1000): Maximum price for listings.
• --no-headless (optional): Use this flag to show the browser window during scraping (disable headless mode). By default, it runs silently in the background.
• --debug (optional): Use this flag to print all raw discovered Marketplace URLs for debugging purposes. This is also useful for seeing which listings are filtered out by the relevance check.

Relevance Filtering:

The scraper now includes a relevance filter. Only listings whose names contain all keywords from your search query (case-insensitive) will be included in the final output. If --debug is enabled, it will show which listings are being filtered out.

Examples:

• Search for "iPhone XR" in "Richmond" with a max price of $500, showing debug URLs:

  python3 main.py --city "Richmond" --query "iPhone XR" --max_price 500 --debug
  

• Search for "Gaming PC" in "Burnaby", showing the browser window:

  python3 main.py --city "Burnaby" --query "Gaming PC" --no-headless
  

Supported Cities:

The main.py script includes a predefined dictionary of cities, primarily focused on Metro Vancouver and surrounding areas in British Columbia. These cities use special Marketplace IDs for improved reliability. You can customize the cities dictionary within main.py to add or modify supported locations:

cities = {
    'Vancouver': 'vancouver', 'Victoria': 'victoria', 'Burnaby': '110574778966847', 'Richmond': '112202378796934',
    'Surrey': '109571329060695', 'Kelowna': '111949595490847', 'Abbotsford': '112008808810771', 'Nanaimo': 'nanaimo',
    'Kamloops': '114995818516099', 'Prince George': '114995818516099', 'Coquitlam': '110019705694079',
    'Langley': '105471749485955', 'Delta': '106083456090237', 'Maple Ridge': '103746596331495', 'New Westminster': '115381681808519',
    'Courtenay': '103113716395848',
}

Output Format:

The scraper prints a JSON array to standard output, where each dictionary represents a listing with the following fields: name, price, location, image (URL), and link (to the listing).

---