Skip to main content

Facebook Marketplace Setup (Optional)

The Facebook Marketplace scraper uses Playwright to browse FB Marketplace listings. This is entirely optional and carries Terms of Service risk — use a disposable Facebook account.

Facebook scraping is disabled by default. You can toggle it on in the Settings page once configured.

Which Fork to Use​

We use a forked version with two new scripts added:

URL
Fork (use this)https://github.com/catesandrew/Facebook-Marketplace_Scraper
Originalhttps://github.com/kevmaindev/Facebook-Marketplace_Scraper

What We Added and Why​

1. capture_session.py​

Opens a visible (non-headless) browser so you can log into Facebook manually, solve any CAPTCHAs, then saves the session cookies to disk for reuse by the headless scraper.

Why: The original script's automated login was failing due to Facebook's aggressive bot detection. Manual login via a visible browser bypasses this entirely. The saved session is then reused for headless searches without triggering login flows again.

2. search_for_app.py​

A headless CLI wrapper that:

  • Accepts command-line arguments: zip code, price range, makes/models list, radius
  • Outputs JSON to stdout (instead of writing CSV files) so the dashboard can consume it directly
  • Filters to trucks and SUVs only via the topLevelVehicleType parameter
  • Uses exact=true and a configurable search radius
  • Extracts listings via stable a[href*="/marketplace/item/"] selectors instead of brittle CSS class names that Facebook changes frequently

Why: The original script wrote CSV output files and used CSS class-based selectors that break whenever Facebook deploys a new frontend. Our version outputs JSON to stdout for direct consumption by the Node.js scraper and uses attribute-based selectors that survive Facebook redesigns.

Setup Steps​

1. Clone the Fork​

git clone https://github.com/catesandrew/Facebook-Marketplace_Scraper.git
cd Facebook-Marketplace_Scraper

2. Install Python Dependencies​

pip install playwright beautifulsoup4 pandas python-dotenv lxml

3. Install Playwright's Chromium Browser​

The NODE_OPTIONS="" prefix is required because Next.js sets --experimental-strip-types in NODE_OPTIONS, which Playwright's bundled Node binary doesn't support.

NODE_OPTIONS="" python3 -m playwright install chromium

4. Capture a Facebook Session​

Run the session capture script with your Facebook email:

NODE_OPTIONS="" python3 capture_session.py your-facebook-email@example.com

A visible Chrome window will open. Log in to Facebook manually, solve any CAPTCHA or security check, then come back to your terminal and press Enter. The script saves your session cookies to a local file.

:::tip Use a Disposable Account Create a throwaway Facebook account for this. Do not use your personal account — if Facebook detects scraping activity, the account may be restricted. :::

Set FB_EMAIL in your car-search/.env.local:

FB_EMAIL=your-facebook-email@example.com

The car-search app passes this to the Python scraper to locate the correct session cookie file.

6. Enable in the Dashboard​

Go to Settings in the dashboard and toggle Facebook Marketplace on. The next scheduled scrape (or a manual Scan Now) will include Facebook listings.

Important Notes​

Session Expiry​

Facebook session cookies expire after some time (days to weeks). When FB listings stop appearing, re-run capture_session.py to refresh the session:

NODE_OPTIONS="" python3 capture_session.py your-facebook-email@example.com

NODE_OPTIONS Prefix​

Always use NODE_OPTIONS="" before Python/Playwright commands when running inside a shell that has the Next.js NODE_OPTIONS set. Without it, you'll see an error like:

error: unknown option '--experimental-strip-types'

This is because Next.js injects --experimental-strip-types into NODE_OPTIONS for TypeScript support, and Playwright's internal Node binary doesn't recognize that flag.

Listing Deduplication​

Facebook listings are deduplicated by their listing URL (with tracking parameters stripped). If the same listing appears across multiple search zips, it's only stored once — lastSeenAt is updated on subsequent scrapes.

ToS Risk​

Using automated tools to scrape Facebook Marketplace violates Facebook's Terms of Service. This scraper is provided for personal/educational use. The risk of account suspension is real — use a disposable account.