Lazada Scraper - Products, Prices, Reviews & Seller Data

Scrape Lazada product data across all six markets - Singapore, Malaysia, Philippines, Indonesia, Thailand and Vietnam - by keyword, category, product URL or item ID. Every product record comes back with its price in the market's own currency, the complete variant tree, the full 5-to-1 star histogram, review counts and, optionally, the newest customer reviews nested right inside the row.

No login. No browser. No CAPTCHA solving. Pure HTTP against Lazada's own server-rendered product pages and its public review API, so runs are fast and cheap.

---

What you get

Product record (search / product / category modes)

Field	Notes
title, itemId, skuId, url	Canonical Lazada product identity
price, priceText, currency	In the market's own currency (SGD/MYR/PHP/IDR/THB/VND)
originalPrice, discountText, discountPercent	Lazada does not server-render these - see the note below the table
brand, brandId, brandUrl	Links to the brand's Lazada page
categoryPath, categoryId, regionalCategoryId	Full breadcrumb path plus Lazada's own category ids
variantOptions	Every option (Colour, Size, ...), its values, and each value's swatch image
skus, skuCount	Per-variant sku id and its own product URL
images, mainImage, imageCount	Every gallery image across all variants
videos	Video id + poster frame - only a minority of listings have one; Lazada publishes no playable URL
description, descriptionHtml, highlights	Clean text and the original HTML
inStock, availability, buyLabel, maxOrderQuantity	Live buyability
ratingAverage, ratingCount, reviewCount	From Lazada's own review API
ratingHistogram, ratingPercent	Exact 5★/4★/3★/2★/1★ counts and percentages
reviewsWithImages, reviewsWithVideos, reviewMediaCount	How much photo/video proof a listing has
topReviews	The newest reviews, nested (rating, text, buyer, date, variant bought)
sellerId	Numeric seller id - Lazada does not publish the seller's name; see below

Three fields Lazada itself withholds - stated up front so nothing surprises you.

1. Was-price / discount %. Lazada server-renders only the current selling price. Measured on 8 products pulled from Lazada's own Choice and Free-Shipping shelves (the ones most likely to be on promotion): the discount field came back empty on 8 of 8, and each page carried exactly one money string. The struck-through price and the -N% badge are drawn client-side after the page hydrates, so there is nothing on the page to read. You still get the exact current price on 100% of rows.
2. Seller name and shop URL. Lazada ships seller_name blank to everyone - measured empty on ordinary marketplace listings and on LazMall (official-store) listings, absent from the review API, and there is no shop link in the page. /shop/{sellerId} is not a valid Lazada URL. Rather than ship a permanently empty column or a broken link, this actor returns the numeric sellerId on 100% of rows, which is stable and groups a seller's whole catalogue.
3. Product video URL. Gallery video entries carry a videoID and a poster frame but no playable link, so that is exactly what you get - no fabricated URLs.

Everything else measured 100% populated on a real 8-product run: title, price, currency, brand, category path, images, variant tree, sku list, description, stock, and the whole rating block.

Review record (reviews mode)

rating · content · ratedAspects (the per-aspect scores buyers fill in, e.g. Sound Quality: 10/10) · buyerName · buyerAvatar · boughtDate · reviewDate · variantBought · skuId · verifiedPurchase · likeCount · upVotes / downVotes · images · videoUrl + videoCover · sellerReply · productTitle · productUrl · isCrossBorder

Reviews are paged 20 at a time straight from Lazada's public review API. Photos (measured 10/21), videos (3/21), seller replies (5/21) and avatars (3/21) exist only where that buyer or seller actually provided them - the counts match Lazada's own summary block exactly, so nothing is being missed.

---

Modes

Mode	What it does
search	Find products by keyword across Lazada's catalogue
product	Scrape specific product URLs or bare item IDs
category	Browse a category by name
reviews	Pull customer reviews for the given products

One mode returns exactly one record type, so a CSV export never columnises on the union of four different shapes and leave you with mostly-blank cells.

How search works (and what it is honestly not)

Lazada walls its on-site search JSON, so this actor searches Lazada's own published product sitemaps instead - up to ~616,000 live products per market, each carrying its full title in the URL. Keyword matching runs over those titles and results are ranked by match quality.

That means: results are ranked by keyword match against Lazada's catalogue, not by Lazada's internal search-relevance ranking. For price monitoring, catalogue building, competitor tracking, assortment analysis and review mining - the things people actually buy this data for - that is what you want. If you need Lazada's exact on-site result order for a query, this is not that.

You can choose which catalogue to search with the catalogue option: all (default, freshest - measured 9/10 URLs still live), lazmall (official stores, 10/10 live), choice, freeShipping, trending, or selling (Lazada's "ordered in the last 30 days" list - demand-ranked, but Lazada regenerates it on a lag so roughly 40% of its entries are already delisted).

---

Example input

{
  "mode": "search",
  "market": "SG",
  "searchTerms": ["wireless earbuds"],
  "catalogue": "all",
  "maxItems": 50,
  "includeReviews": true,
  "maxReviewsPerProduct": 5,
  "minReviews": 10,
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Filter as you go with minPrice, maxPrice, minRating and minReviews - these run inside the crawl, so maxItems returns that many matching products rather than being eaten by rows that get thrown away.

Example output (trimmed)

{
  "type": "product",
  "market": "SG",
  "itemId": "3427720191",
  "title": "Vention IPX4 Bluetooth 5.4 Wireless Earbuds ... Noise Cancelling Earphone",
  "price": 32.99,
  "currency": "SGD",
  "brand": "Vention",
  "categoryPath": ["Audio", "Headphones & Headsets"],
  "inStock": true,
  "skuCount": 3,
  "variantOptions": [
    { "name": "Color Family", "values": [
      { "name": "White", "image": "https://laz-img-sg.alicdn.com/p/c03f...jpg" },
      { "name": "Black", "image": "https://laz-img-sg.alicdn.com/p/34c9...jpg" }
    ]}
  ],
  "ratingAverage": 4.8,
  "reviewCount": 213,
  "ratingHistogram": { "star5": 190, "star4": 15, "star3": 5, "star2": 1, "star1": 2 },
  "sellerId": "1000061401"
}

---

Reliability

Lazada rate-limits by IP: the same product page that is throttled on one exit returns the full document on another. This actor is built around that measured fact - each parallel worker owns its own residential proxy session (its own IP), sessions are retired after a few requests, and every retry rotates to a fresh IP rather than re-hitting the throttled one. Measured 4/8 success on a single cold attempt, which the rotation takes to ~98% per product.

A throttle is never treated as "this product has no data", and it never crashes the run: blocked runs finish Succeeded with a status message telling you to re-run. The actor also reads its own run deadline and stops cleanly before the platform can kill it, so a big job ends with partial results instead of a TIMED-OUT run.

Use with AI assistants (MCP)

This actor works as a tool for AI assistants through the Apify MCP server - point Claude, ChatGPT or your own agent at it and ask for Lazada product, price or review data in plain language.

Agent-ready: autonomous payments (x402 & Skyfire)

This actor is agent-ready — AI agents can discover it, run it, and pay for it autonomously, with no Apify account and no human in the loop. It uses pay-per-event pricing and limited permissions, so it qualifies for Apify's agentic-payment standards:

• x402 — an open, HTTP-native payment protocol. Agents pay per run in USDC on the Base network directly through the Apify MCP server — no account, no API key.
• Skyfire — agent-to-service payments for fully autonomous AI-agent workflows.

Building an AI agent, MCP tool, or autonomous data pipeline? This scraper is ready to plug in and pay as it goes.

---

Pricing

Pay per result, no start fee:

• Product scraped - $0.004 - the complete record above, product page + review API in one pass.
• Review scraped - $0.001 - one full customer review.

Legal & fair use

This actor collects publicly visible product information - the same pages Lazada serves to any visitor and publishes in its own sitemaps for search engines. It does not log in, does not touch cart, checkout or account pages (which Lazada's robots.txt is the only thing to disallow, and which this actor never requests), and collects no personal data beyond the public display name a reviewer chose to show on their own public review. You are responsible for using the data in line with Lazada's terms and applicable law, including data-protection rules where reviewer names are involved.