MercadoLibre Reviews Scraper

Scrape product reviews and rating summaries from MercadoLibre and MercadoLivre across 18 Latin-American countries. For every catalog product you get the full review list plus the rating summary MercadoLibre computes itself: average score, the 5→1 star distribution, photo/video counts, and an AI-generated review summary.

Unlike other review scrapers, this one returns the AI summary and star distribution — not just raw review text. And it runs against an open endpoint, so no proxy is required.

Who it's for

• Brands & product researchers — read one AI summary and a star breakdown instead of 800 individual reviews.
• Sellers — track review velocity, complaints and photo evidence on your own and competitors' catalog products.
• Analysts & ML teams — assemble clean, labelled review datasets (rating, date, media, country) for sentiment analysis.

Output

The dataset contains two row types, distinguished by record_type.

summary — one row per product

Field	Type	Description
product_id	string	Catalog product id
rating	number	Average score (0–5)
total_ratings	number	Total number of ratings
total_comments	string	Comment count label
stars_5 … stars_1	number	Count of ratings at each star level
pictures	number	Review photos across the product
videos	number	Review videos across the product
ai_summary	string	MercadoLibre's AI-written review digest
site	string	Country site id
scraped_at	string	ISO-8601 timestamp

review — one row per review

Field	Type	Description
review_id	number	Review id
rating	number	Star rating (1–5)
text	string	Review body
date_relative	string	Relative date (e.g. Hace 9 meses)
created_date	string	Absolute ISO-8601 date
country	string	Reviewer country
variant	string	Purchased variant, when shown
likes	number	Helpful-vote count
picture_count / video_count	number	Media attached to the review
picture_urls	array	Review photo URLs
site	string	Country site id
scraped_at	string	ISO-8601 timestamp

Use cases

• Sentiment at a glance — pull the summary rows for a catalogue and rank products by rating and 1-star share.
• Complaint mining — filter review rows to rating <= 2 and cluster the text.
• Photo evidence — collect picture_urls from high-rated reviews for listings or ads.
• Competitor monitoring — schedule daily and watch total_ratings and the distribution move.

Input parameters

Field	Type	Default	Description
productIds	array	["MLA57692573"]	Catalog product ids or /p/ URLs
site	enum	MLA	Default country site for plain ids
maxReviewsPerProduct	integer	100	Per-product review cap
includeSummary	boolean	true	Emit a summary row per product
maxItems	integer	100000	Global row cap
proxyConfiguration	proxy	off	Optional — the endpoint is open

Example input

{
  "productIds": ["MLA57692573", "https://www.mercadolibre.com.mx/x/p/MLM1466527953"],
  "maxReviewsPerProduct": 200,
  "includeSummary": true
}

Getting product IDs

This actor needs catalog product ids — the MLA…/MLB…/MLM… code in a product URL's …/p/MLA57692573 segment. You can feed the full product URL and the actor extracts the id. (Item ids like MLA-1234567890 are seller listings without a catalog review page.) Pair it with the MercadoLibre Scraper actor, whose rows include product_id, to review an entire search or category.

Scheduling & integrations

Run on a schedule and export to Google Sheets, Make, Zapier or S3, or pull the dataset via API:

https://api.apify.com/v2/datasets/<datasetId>/items?format=json

Use it from an AI assistant (MCP)

Callable from AI agents via the Apify MCP server — an assistant can fetch "the rating summary and worst reviews for product MLA57692573" and read structured rows.

FAQ

Do I need a proxy? No — reviews come from an open endpoint.

Why do I get "no reviews found"? You likely passed an item id (MLA-1234…) instead of a catalog product id (MLA57692573), or the product has no reviews yet.

How many reviews can I get per product? Set maxReviewsPerProduct; the actor pages until it reaches that cap or runs out of reviews.

Which countries? All 18 MercadoLibre sites (Argentina, Brazil, Mexico, Chile, Colombia, Peru, and more).

Is it legal? The actor collects only public review data. You are responsible for using it in line with MercadoLibre's terms and applicable law.