MercadoLibre Official Stores Scraper

Scrapes the public official-stores directory ("tiendas oficiales") of MercadoLibre Mexico and returns one record per official store: brand name, store URL, logo URL, and directory category.

Who would use it: lead-list builders, brand/marketplace analysts, and anyone maintaining a directory of brands with an official storefront on MercadoLibre Mexico. Official store names are business names shown on public listings; no personal data is collected.

Inputs

Field	Type	Default	Description
start_urls	array of strings	["https://www.mercadolibre.com.mx/tiendas-oficiales"]	Directory pages to start from. Accepts the directory landing page (category pages are discovered automatically) or individual category pages (tiendas-oficiales/show?category=MLM1000...).
max_items	integer	250	Maximum number of unique store records to return. The actor visits category pages in order and stops as soon as this many unique stores have been collected.
country	string	"mx"	Two-letter country code stamped on each record.

Sample output record

{
  "store_name": "117 Mx",
  "store_url": "https://www.mercadolibre.com.mx/tienda/117-mx",
  "logo_url": "https://http2.mlstatic.com/D_NQ_NP_832107-MLA112151539782_062026-F.jpg",
  "category": "Electrónica, Audio y Video",
  "country": "mx"
}

Fields: store_name (str), store_url (str), logo_url (str or null), category (str or null), country (str). A store that appears in several categories keeps the category where it was first seen; logo_url prefers the brand's high_resolution_logo asset.

Coverage and limits

• The directory is organized by product category (24 categories such as "Electrónica, Audio y Video", "Ropa, Bolsas y Calzado", "Computación"). Each category page (tiendas-oficiales/show?category=MLM...) renders a fixed set of top brands (100 per page, server-side, with store data also embedded in the page's __NORDIC_RENDERING_CTX__ JSON).
• Pagination is by category, not by page. The category pages ignore page=/letter= query parameters and expose only their first ~100 brands even when the page reports a larger totalBrands count. So coverage is the top ~100 brands per category, up to ~2,400 raw listings deduped by store URL.
• Deliberately avoided surfaces: MercadoLibre search-result pages and product-detail pages redirect unauthenticated visitors to account verification; this actor never requests them. If any directory page starts redirecting to a login/account-verification wall or serves a CAPTCHA, the actor logs the URL, status code, and redirect target and moves on.

Rate limiting

Plain HTTPS with a browser User-Agent, max 2 concurrent connections, and a fixed ~1 second delay between requests. No scraping proxy services, no CAPTCHA solving, no parallel hammering. On hosts that route outbound traffic through a standard egress proxy, the actor uses the machine's configured https_proxy/http_proxy URL explicitly (this host's no_proxy format breaks httpx's environment parsing, so the URL is passed directly rather than relying on trust_env).

Data rules

• Public pages only. The actor visits only the official-stores directory landing and category pages, all accessible without login. No login flows, no session cookies.
• No personal data. Only business/brand names, public store URLs, and logo image URLs are collected. No names, emails, or phone numbers of private individuals.

Run locally

cd mercadolibre-official-stores-scraper
mkdir -p storage/key_value_stores/default
cat > storage/key_value_stores/default/INPUT.json <<'EOF'
{
  "start_urls": ["https://www.mercadolibre.com.mx/tiendas-oficiales"],
  "max_items": 250,
  "country": "mx"
}
EOF
python main.py

Output lands in storage/datasets/default/ (local Apify storage). The same files build the Apify Docker image (Dockerfile, requirements.txt) without changes.