Airbnb Search Scraper

Scrapes Airbnb search results for any location — title, price with discount breakdown, rating, review count, coordinates, images. No login, no API key.

Useful for market research on short-term rental pricing, competitor listing tracking, and building a location-price dataset across many cities.

What makes this different

An unrecognized location doesn't quietly return the wrong thing. Search a place name Airbnb can't resolve, and it doesn't error — it silently substitutes a global "explore" feed full of real-looking listings from random cities worldwide. There's exactly one tell in the response that distinguishes a genuine match from that fallback, and this Actor checks it before returning anything, so a typo'd location fails loudly instead of quietly shipping listings from the wrong place entirely.

Duplicate listings across pages are caught and compensated for, not just counted. Airbnb repeats certain listings — apparently boosted or promoted ones — across multiple pages of the same search; the very first listing on page 1 can reappear verbatim on page 2. A scraper that doesn't dedupe would report "60 results" that are actually 40-some unique listings padded with repeats. This Actor deduplicates by listing ID within a run and pages further to make up the difference, so the count you ask for is the count of distinct listings you get.

The full discount picture, not just a single price. price_display and price_original_display are both kept, along with has_discount, so a promotional price is visible as exactly that — not indistinguishable from a listing's normal rate.

Input

Field	Type	Notes
locations	array	City, region, or "City, Country" — same as Airbnb's own search box.
currency	string	3-letter code, e.g. USD, EUR, IDR.
priceMin / priceMax	integer	Optional nightly price range, in the chosen currency.
maxItems	integer	Per location, deduplicated. 0 walks every page available.
incremental	boolean	Only listings not seen in previous runs.
proxyConfiguration	object	Not normally needed.

Output

{
  "listing_id": "airbnb.com:672988127335230371",
  "title": "Room in Paris",
  "subtitle": "Room with terrace & art exhibition",
  "rating": 4.88,
  "review_count": 151,
  "price_display": "$558",
  "price_original_display": "$617",
  "has_discount": true,
  "badges": ["Guest favorite"],
  "latitude": 48.82888,
  "longitude": 2.32166,
  "url": "https://www.airbnb.com/rooms/672988127335230371"
}

A location Airbnb doesn't recognize fails that location's run with a clear message rather than returning listings from somewhere else entirely. Each run also writes a RUN_COVERAGE record to the key-value store with what was requested, what came back, and any per-location failures.

Local development

pip install -r requirements.txt
python test_local.py Paris --max 40 --out sample_output.json
python test_local.py "Bali, Indonesia" --currency USD --price-min 20 --price-max 100

sample_output.json is real output from a live 40-listing run, deduplicated across pages.