Airbnb Listings Scraper

Collect a bounded set of publicly displayed Airbnb search listings without login. The actor requests the public search page directly, follows Airbnb's ordinary country-domain handoff when present, and extracts the structured listing data returned in that page.

Behavior

• Direct HTTP is always tried first. It does not silently use a proxy.
• Direct HTTP is the only transport. A blocked page returns an uncharged diagnostic and never starts a residential proxy.
• Each location is capped at 20 output rows; at most five locations may be supplied per run.
• The overall wall-clock deadline is capped at 180 seconds.
• A genuine listing is written before its listing result event is charged. Samples, diagnostics, duplicates, empty results, and blocked responses are not charged.
• Empty input returns one explicit sample row and makes no network request or charge.

Input

locations accepts city, region, or destination strings, such as Toronto, Ontario, Canada. Optional checkIn and checkOut must both be ISO dates, with checkout after check-in. Supplying dates lets Airbnb return date-specific displayed prices. adults is bounded from 1 to 16.

Output

Each genuine listing row includes the public listing ID and URL, name, property type, displayed price and currency, rating/review count when available, badges, concise space details, optional map coordinates, up to five public image URLs, search inputs, source, and transport (direct). Prices are displayed search-page values, not a booking quote or availability guarantee.

Limitations

Airbnb can vary, block, or remove public search data by locale and request context. This actor does not log in, bypass CAPTCHA, collect private data, or promise availability. It returns an uncharged diagnostic when public data is unavailable instead of inventing listings or billing empty output.

Example

{
  "locations": ["Toronto, Ontario, Canada"],
  "checkIn": "2026-08-10",
  "checkOut": "2026-08-12",
  "adults": 2,
  "maxItemsPerLocation": 20,
  "deadlineSecs": 90
}