MercadoLibre Scraper

Scrape products from MercadoLibre — Latin America's largest e-commerce marketplace. Returns product ID, title, URL, price, currency, condition, seller, thumbnail, rating, and more. Supports Argentina, Brazil, and Mexico.

What it does

• Scrapes MercadoLibre search results across 3 LatAm country sites (AR, BR, MX)
• Extracts product ID, title, URL, price, currency, thumbnail, rating, reviews, seller, brand, and shipping info
• Hardcoded residential proxy handles MercadoLibre's account-verification block wall
• Automatic session rotation — up to 6 retries on flagged sessions
• Clean, non-null output — empty fields are omitted from records

Input

Field	Type	Default	Description
country	enum	AR	Which MercadoLibre site to scrape: AR (Argentina), BR (Brazil), MX (Mexico)
searchQuery	string	"iphone"	Free-text keyword (required unless startUrls is set)
condition	enum	any	any / new / used
priceMin	integer	—	Minimum price in local currency
priceMax	integer	—	Maximum price in local currency
startUrls	array	—	Direct MercadoLibre search URLs (override structured filters)
maxItems	integer	50	Maximum results to return (1–1000)

Example input

{
    "country": "BR",
    "searchQuery": "notebook",
    "priceMin": 2000,
    "priceMax": 10000,
    "maxItems": 20
}

Output

Each product is a JSON record with the following fields (missing fields are omitted, never null):

Field	Type	Always present	Description
id	string	✓	MercadoLibre item ID (e.g. MLA1018500855)
title	string	✓	Product title
url	string	✓	Full product URL
countryCode	string	✓	Country code (e.g. AR)
scrapedAt	string	✓	UTC ISO 8601 scrape timestamp
price	number	—	Current price in local currency
originalPrice	number	—	Strikethrough / original price (before discount)
currency	string	—	Currency code (e.g. ARS, BRL, MXN)
currencySymbol	string	—	Currency symbol as displayed on the site
discountPercentage	number	—	Percentage off (0–100)
thumbnail	string	—	Product image URL
sellerName	string	—	Seller / brand name
brand	string	—	Brand attribute
shippingLabel	string	—	Shipping badge text (e.g. Envío gratis, Frete grátis)
rating	number	—	Seller or product star rating
reviewsCount	number	—	Number of reviews

Sample output

{
    "id": "MLA1018500855",
    "title": "iPhone 13 128 GB 4 GB RAM Blanco estelar Apple - Distribuidor Autorizado",
    "url": "https://articulo.mercadolibre.com.ar/MLA-1018500855-iphone-13-128-gb-...",
    "countryCode": "AR",
    "price": 999999,
    "currency": "ARS",
    "currencySymbol": "$",
    "thumbnail": "https://http2.mlstatic.com/D_Q_NP_2X_787642-MLA95698363690_102025-E.webp",
    "sellerName": "DistribuidorOficial",
    "shippingLabel": "Envío gratis",
    "scrapedAt": "2026-04-14T06:38:23Z"
}

FAQ

Do I need a proxy? No configuration needed — an Apify residential proxy is applied automatically, matching the selected country. MercadoLibre blocks direct datacenter access and even many residential sessions, so the scraper rotates proxy session IDs up to 6 times per run on flagged sessions. The first half of attempts use the target country's residential pool; the second half falls back to any residential IP for resilience.

Why only 3 countries? At the time of writing, MercadoLibre's smaller-market sites (Chile, Colombia, Peru, Uruguay, Venezuela, Ecuador, Dominican Republic, Costa Rica) return a hard-block response on every residential session in Apify's pool. Only AR, BR, and MX reliably serve scrapable HTML. The scraper's country enum has been trimmed accordingly.

How many results can I get per run? Up to 1,000 items (theoretically). Each search page returns 40–50 cards; the scraper paginates via the MercadoLibre page index when maxItems exceeds a single page. Practical cap depends on how deep the search goes before MercadoLibre starts returning "no more results".

Does the scraper call MercadoLibre's public API? No. api.mercadolibre.com now returns 403 for anonymous traffic. The scraper uses the HTML search listings page (listado.mercadolibre.com.ar/<query>, etc.) and parses the poly-card server-rendered product tiles.

What happens if the target site is blocked? The scraper retries with a fresh residential session. If all 6 retries fail, the run exits with a fail status and no records — the residential pool was exhausted for that country. Rerun in a few minutes.

How fresh is the data? Every run hits the live search page. Prices and availability are always current.

Use cases

• Price monitoring — daily runs to track competitor prices in AR/BR/MX
• Trend research — ingest product names and prices across major keywords
• Cross-border arbitrage — compare pricing between AR, BR, and MX markets
• Inventory alerts — monitor specific SKUs via the startUrls override

Notes

• Pricing is configured in the Apify UI (pay-per-result).
• MercadoLibre's HTML can change; if the scraper starts returning 0 items, check the poly-card class markers in src/main.py:parse_search_html.
• The daily Apify test run uses country=AR, searchQuery=iphone, maxItems=5.