Falabella Product Search Scraper

Apify actor that scrapes public, logged-out product search results from falabella.com (Colombia default; Chile, Peru, and Argentina storefronts also supported).

What it scrapes

For a given search query, the actor walks the search-results pages (https://www.falabella.com.co/falabella-co/search?Ntt=<query>&page=N) and extracts one record per product listing:

Field	Type	Description
title	string	Product display name
brand	string or null	Brand name as shown on the listing
price_cop	float or null	Current price shown (lowest non-crossed-out price; COP on the Colombia storefront)
original_price_cop	float or null	Crossed-out reference price when a discount is shown
product_url	string	Canonical product page URL
image_url	string or null	First product image URL (Falabella CDN)
rating	float or null	Average rating (e.g. 4.57) when reviews exist

Falabella server-renders search pages as Next.js, so the actor parses the embedded __NEXT_DATA__ JSON with httpx + beautifulsoup4/lxml — no JavaScript engine, no headless browser, no login, no session cookies beyond the anonymous ones the site sets itself.

Who would use it

Price-comparison builders, retail analysts, and marketplace researchers who need structured Falabella catalog data (titles, brands, prices, ratings) for a keyword without running a browser farm.

Input

Field	Type	Default	Description
query	string	"audifonos bluetooth"	Product search query (required).
country	string	"co"	Storefront country: co (Colombia), cl (Chile), pe (Peru), ar (Argentina).
max_items	integer	60	Maximum product records to scrape. The actor paginates (&page=2, &page=3, ...) until this many records are collected or results run out.

Sample output record

{
  "title": "Audífonos bluetooth WH-CH520",
  "brand": "SONY",
  "price_cop": 119900.0,
  "original_price_cop": 299900.0,
  "product_url": "https://www.falabella.com.co/falabella-co/product/prod13360109/Audifonos-bluetooth-Sony-WH-CH520",
  "image_url": "https://media.falabella.com.co/falabellaCO/73319593_1/public",
  "rating": 4.6594
}

(Real record from a 2026-09-26 test run against the Colombia storefront.)

Coverage and limits

• Search results are paginated server-side; the actor stops at max_items or when the result set is exhausted (the site reported ~1,415 results / 30 pages for the test query).
• price_cop reflects the lowest current listed price; event/promo prices (eventPrice) and card prices (cmrPrice) are treated as current prices.
• Nulls are normal: unrated products have rating: null, and listings without a crossed-out price have original_price_cop: null.
• One homepage request seeds anonymous cookies, then ~1 request/second between result pages (politeness throttle).
• Only the search?Ntt= surface is scraped. Falabella redirects the legacy ?text= parameter to the homepage, so the actor uses Ntt (the parameter in the site's own schema.org SearchAction).
• Field names say price_cop; on non-Colombia storefronts the price is in that country's listing currency.

Data rules

• Public, logged-out pages only. No login flows, no stored session cookies, no CAPTCHA solving, no proxy services, no headless browsers (plain HTTP).
• Product and brand names on public listings are fair game; the actor never collects anything identifying private individuals.
• Respects a ~1 request/second pace between pages and stops cleanly after three consecutive page failures.

Local test

# write input
mkdir -p storage/key_value_stores/default
printf '{"country":"co","max_items":60,"query":"audifonos bluetooth"}' \
  > storage/key_value_stores/default/INPUT.json
# run with the shared venv
/path/to/.venv/bin/python main.py
# results land in storage/datasets/default/