QuintoAndar Property Scraper (Brazil)

Scrapes apartments and houses for rent or sale from QuintoAndar — Brazil's largest rental-focused real estate platform.

Public data only. No login, no cookies, no browser.

The one thing you need to know before using this

Each location returns a single fixed batch of 12-26 listings — there is no pagination this actor can reach. QuintoAndar's search page always renders the exact same first batch server-side regardless of any page/sort/ offset parameter (all tried and confirmed to change nothing); real pagination happens client-side against a /busca endpoint that robots.txt explicitly disallows for generic crawlers. This actor honours that. locations being a list is therefore the only way to cover more ground — query many neighbourhoods instead of one city, the way you would split a search into narrower slices anyway.

What you get

Three record types share one dataset, told apart by recordType.

PROPERTY — one row per listing

Search rows (listing) carry price, area, bedrooms, bathrooms, neighbourhood, condo fee (IPTU), parking spots, amenities and photos. With Fetch full property details on (the default), each row also gets a propertyDetails object with the detail page's richer payload: a full boolean-flag amenity list, exact street address, and the site's internal display id.

SEARCH_SUMMARY — one row per location searched

Upstream's own claimed total (informational only — see above, this actor can never reach more than the fixed batch) and how many rows this run actually returned.

Input

Field	What it does
Alugar / comprar	rent or buy
Locations	QuintoAndar slugs: sao-paulo-sp-brasil, vila-madalena-sao-paulo-sp-brasil… one fixed batch per entry
Property URLs	scrape specific listings directly, skipping search
Fetch full property details	off = fast listing-only crawl

Example

{
  "transactionType": "alugar",
  "locations": ["vila-madalena-sao-paulo-sp-brasil", "pinheiros-sao-paulo-sp-brasil", "itaim-bibi-sao-paulo-sp-brasil"],
  "includePropertyDetails": true
}

Notes on reliability

• No WAF challenge. Serves data cold on every TLS profile tested.
• Runs unattended. No human-captured session, no expiring token.
• Unknown locations fail cleanly as not_found — QuintoAndar answers with an explicit houses.error: true, never a silent fallback to a nationwide or default search (verified live; a genuine exception among real-estate targets in this portfolio, most of which DO silently widen on a bad filter).
• Residential proxy by default on cloud runs.
• Failures never vanish. A listing withdrawn mid-crawl degrades that one row (_detailError); every input maps to at least one output row.

Output envelope

Every record carries _input, _source and _scrapedAt. Upstream field names pass through verbatim under listing (search) and propertyDetails (detail) — no renaming. _source is S1-nextdata-search for listing-only rows and S1-nextdata-search+S2-nextdata-detail once details are attached.

See CRAWLING_METHOD.md for the full reverse-engineering trail, including why pagination isn't offered and a sample of QuintoAndar's ~99 supported city slugs.