Lazada Reviews Scraper — Ratings, Media & Seller Intel

Scrape Lazada product reviews as clean JSON: review text, per-aspect scores, verified-purchase flags, buyer photos and videos — plus the complete rating breakdown of the product itself. Works on all six Lazada marketplaces (.vn, .co.id, .com.my, .sg, .co.th, .com.ph). API-level, no brittle browser automation.

What you get

Give it Lazada product URLs. The dataset contains two kinds of item, tagged by a type field. Products and reviews are separate rows, joined by itemId.

Product — emitted once per product, and it is not just an echo of the page:

{
  "type": "product",
  "itemId": 246452966,
  "title": "Newborn nappy pads, 108 pcs",
  "priceRaw": "139.000 ₫",
  "sellerId": 100101649,
  "subSellerId": 1000001034,
  "categoryId": 10614,
  "categoryName": "Tã giấy",
  "average": 4.9,
  "rateCount": 1243,
  "reviewCount": 1243,
  "stars5": 1221,
  "stars4": 6,
  "stars3": 2,
  "stars2": 3,
  "stars1": 11,
  "negativeCount": 14,
  "hiddenCount": 864,
  "withImageCount": 151,
  "withVideoCount": 4,
  "withMediaCount": 154,
  "url": "https://www.lazada.vn/products/...-i246452966-s316699339.html"
}

The full star histogram (stars5…stars1), the count of unhappy buyers (negativeCount = 1★+2★) and the media counts arrive in the same response as the reviews, so you get product-level sentiment without paying for a second pass. priceRaw is deliberately a string in the marketplace's own format ("139.000 ₫") — we do not guess a decimal separator and hand you a wrong number.

Review:

{
  "type": "review",
  "itemId": 246452966,
  "reviewId": "464679055652966",
  "rating": 5,
  "ratingProduct": 5,
  "ratingSeller": 5,
  "ratingLogistics": 5,
  "text": "Giao hàng nhanh, đóng gói cẩn thận",
  "author": "8***9",
  "isAnonymous": false,
  "verifiedPurchase": true,
  "skuInfo": "Kích thước tã:Trẻ sơ sinh (Tới 5kg), Kích thước gói tã:108",
  "skuId": "316699339",
  "reviewTimeRaw": "08 thg 6 2026",
  "reviewDate": "2026-06-08",
  "boughtDateRaw": "20 thg 6 2026",
  "images": ["https://lzd-u.slatic.net/....jpg"],
  "videoUrl": "https://lazvideo.lazcdn.com/....mp4",
  "likeCount": 0,
  "upVotes": 0,
  "downVotes": 0,
  "sellerId": 100101649,
  "url": "https://www.lazada.vn/products/...-i246452966.html"
}

Lazada splits a review into three separate scores — product, seller and delivery — and this scraper keeps all three (ratingProduct, ratingSeller, ratingLogistics) alongside the overall rating. verifiedPurchase is the flag Lazada's own front end uses to draw its "Verified Purchase" badge. sellerReply and followUpText appear when the store answered the review or the buyer added a follow-up. Photos come back as a flat list of URLs; videoUrl is a signed link and expires, so download it promptly if you need the file.

Use cases

• Review mining & sentiment analysis for brands selling in Southeast Asia. Lazada keeps three separate scores per review — product, seller, delivery — so you can tell a product complaint from a courier complaint instead of averaging them into mush.
• Seller reputation monitoring. The star histogram, the 1–2 star count and the seller's public replies (sellerReply) per product show how a store handles unhappy buyers over time.
• Collecting buyer photos and videos — real-world product shots for catalog QA or content research. Photo URLs come as a flat list; video links are signed and expire, so download promptly.
• Competitor research before entering Lazada: rating breakdowns, review volume and category data across competing listings on any of the six marketplaces.

Input

Field	Type	Description
productUrls	string[]	Lazada product page URLs. The product id and the marketplace domain are read from the link itself.
maxReviews	integer	How many reviews to keep per product (default 50).
sort	integer	0 relevance, 1 most recent (default), 2 rating high→low, 3 rating low→high.
filter	integer	0 all reviews, or 1–5 for exactly that many stars.
proxyConfiguration	object	Proxy to route requests through. Defaults to residential Vietnam — see "Limits" below before changing it.

Why the default sort is "most recent", not Lazada's own default. On a measured product, relevance ordering put fifty 5-star reviews on page one and pushed every 1–2 star review to the last page. An actor that grabs "the first N reviews" under that ordering quietly hands you a positively biased sample. This one defaults to chronological order; set sort: 0 if you specifically want Lazada's relevance ranking.

How to use it

1. Click Try for free / Start on this page.
2. Paste one or more Lazada product page URLs into Product URLs. Any of the six marketplaces works — the product id and the marketplace domain are read straight from the link, and you can mix domains in one run.
3. Optionally set Max reviews per product, the Sort order and the Star filter (for example, 1 to pull only 1-star reviews for complaint mining).
4. If your URLs are not from lazada.vn, change the country in Proxy configuration to match the marketplace (ID, MY, SG, TH, PH) — see "Limits" below for why this matters.
5. Run it. Watch the Output tab: the Products view shows one row per product with the full rating breakdown, the Reviews view shows the review rows. Export as JSON, CSV, Excel or pull the data over the API.

To run it on a schedule, use Apify Schedules; to call it from your own code, use the Apify API or one of the official clients. Every field in the output is described in this Actor's dataset schema, so AI agents calling it through the Apify MCP server get typed field metadata rather than raw JSON.

Limits (honest)

• You cannot get every review of a product, and no scraper can. Lazada's public review endpoint serves only about rateCount − hiddenCount reviews; the rest are withheld by Lazada itself and are unreachable at any page number. On the product we measured, that was 379 of 1243 reviews, i.e. roughly 30%. Both numbers ship in the product row (rateCount and hiddenCount), so you can always see how much of the corpus you received rather than guessing.
• Lazada's own page counter overstates the supply. Its pagination metadata advertised 25 pages where only 8 held data. The scraper stops when the data actually runs out, not when Lazada claims it has.
• Machine-readable review dates do not exist in this response. Lazada returns localized strings that mix absolute dates ("08 thg 6 2026") with relative ones ("2 tuần trước" — "2 weeks ago"). The raw string is always preserved in reviewTimeRaw / boughtDateRaw. reviewDate is filled in only when the raw value parses unambiguously; when it does not, the field is simply absent. We do not substitute today's date for a value we could not read.
• The seller's name is not in this endpoint. Lazada returns null for it. You get numeric sellerId / subSellerId and the category, which is what is actually there.
• Rate limits are strict, and the proxy country matters more than the pace. Lazada answers with a challenge page when it does not like the caller, and it dislikes datacenter IPs in particular: from one, the endpoint served roughly a single clean response every five minutes. From residential IPs in the marketplace's own country it sustained about ninety times that. The default proxy setting is therefore residential Vietnam — change the country if you scrape another marketplace. The scraper fetches one page at a time, and treats a challenge as a failed task to be retried on a fresh IP, never as "the reviews ended".
• Public data only: no login, no captcha solving, no circumvention of access controls.

Pricing

Pay-per-event: a charge per run start, a charge per scraped product, and a separate, cheaper charge per scraped review, so a run over many products with few reviews is not billed as if it were review-heavy. See the Store page for current rates.

FAQ

Do I need a Lazada account or an API key? No. The Actor reads only publicly available data, never logs in and does not solve captchas or circumvent access controls.

Which Lazada marketplaces are supported? All six: lazada.vn, lazada.co.id, lazada.com.my, lazada.sg, lazada.co.th and lazada.com.ph. The marketplace is detected from each URL's domain, so one run can mix them.

Why did I get fewer reviews than the product page shows? Because Lazada itself withholds part of the corpus. Its public review endpoint serves only about rateCount − hiddenCount reviews — on the product we measured, 379 of 1243. Both numbers ship in the product row, so you can always see exactly how much of the corpus you received. No scraper can get past this; one that claims otherwise is counting duplicates.

Why is the default proxy residential Vietnam? Because that is what the endpoint tolerates. Measured on the same code: a datacenter IP got roughly one clean response every five minutes, while Vietnamese residential IPs sustained about ninety times that. Vietnam is the default because lazada.vn is this Actor's primary market — if you scrape another marketplace, switch the proxy country to match it.

Why do some reviews have no reviewDate? Lazada returns localized date strings that mix absolute dates with relative ones ("2 weeks ago"). reviewDate is filled in only when the raw value parses unambiguously; the raw string is always preserved in reviewTimeRaw. We do not substitute today's date for a value we could not read.

Why does the review video link stop working? videoUrl is a signed link from Lazada's CDN and expires. Download the file promptly if you need it; the review row itself stays in your dataset.

Other Actors by ActorForge

• Wildberries Scraper — products, prices and reviews from Wildberries as clean, schema-validated JSON.
• Avito Real Estate Scraper — full property inventory from Avito with freshness stamps, for new-listing alerts and comps.

Need another Lazada region, seller-level data, or a different marketplace? Open an issue from the Actor's page and tell us what you need.

Disclaimer

This is an unofficial scraper. It is not affiliated with, endorsed by, or connected to Lazada, Lazada Group or Alibaba Group in any way. It collects only data that Lazada publishes publicly, without logging in and without circumventing access controls. Review text and buyer media belong to their authors; you are responsible for ensuring your use of the collected data complies with applicable law and with Lazada's terms.

Changelog

See the repository CHANGELOG.md.