TikTok Shop Recommendations Scraper

The Actor collects one of three recommendation carousels that TikTok Shop attaches to a product page: the you may like rail, the more from this shop rail, or the seller's top-reviewed products. scraperType selects which carousel a run collects, and the two rail families return different card shapes rather than one shared shape.

{
  "scraperType": "youMayLike",
  "product_ids": [
    "1730927783781307026",
    "https://shop.tiktok.com/us/pdp/maybelline-new-york-serum-lipstick-8h-moisture-satin-matte/1731756086010286738"
  ],
  "region": "US",
  "maxItems": 100,
  "enrichProductDetails": false
}

Accepted input

scraperType is required and defaults to youMayLike; the other options are moreFromShop and topReviewed. product_ids is required and accepts TikTok Shop product page links, product IDs, or both mixed in the same list; each product is fetched with its own request, so a run against ten product_ids makes ten requests regardless of scraperType. region is required, defaults to US, and selects the storefront read (US, GB, SG, MY, PH, TH, or VN); this changes the products, prices, and currency returned, not only the display language. maxItems defaults to 100 and bounds the number of rows saved across the whole run; a value of 0 removes the bound. enrichProductDetails defaults to false and adds the full product page to every row: description, every variant, the shop profile, a page of reviews with the star breakdown, and the category path. It costs one extra request per row, so it is worth leaving off on a first pass.

Result fields

Every row carries product_id, rail (you_may_like, more_from_shop, or top_reviewed, matching scraperType), and source_product_id, which names the product whose page produced the recommendation rather than the recommended product itself. Beyond those three, the row shape depends on which carousel was collected.

youMayLike and moreFromShop return TikTok's standard listing card: title, image, product_price_info, rate_info, sold_info, seller_info, seo_url, sku_info, brand_info, and product_marketing_info. topReviewed returns a different card built around review standing rather than a listing feed: name where the listing card has title, price_info where it has product_price_info, product_rating and review_count where it has rate_info, review_overviews (short excerpts pulled from real reviews), and canonical_url where it has seo_url. A youMayLike or moreFromShop row leaves name, price_info, product_rating, review_overviews, and canonical_url null; a topReviewed row leaves title, product_price_info, rate_info, sold_info, and seo_url null. product_card holds the untouched card exactly as TikTok returned it either way, so nothing broken out above is ever the only copy.

With enrichProductDetails set to true, every row also gains shop_info, review_info, categories, promotion_tag, and product_detail: the same detail fields the product's own page carries.

{
  "product_id": "1731896769144001226",
  "rail": "you_may_like",
  "source_product_id": "1730927783781307026",
  "title": "Hydrating Lip Gloss - High Shine & Tinted, Infused with Hyaluronic Acid, Jojoba Oil & Shea Butter",
  "product_price_info": { "currency_name": "USD", "sale_price_format": "13.23", "discount_format": "59%" },
  "rate_info": { "score": 4.5, "review_count": "7814" },
  "sold_info": { "sold_count": 73428 },
  "seller_info": { "seller_id": "7495153913553128138", "shop_name": "Truly Beauty" }
}

The same page's topReviewed carousel returns a card shaped like this instead:

{
  "product_id": "1729447567685489298",
  "rail": "top_reviewed",
  "source_product_id": "1730927783781307026",
  "name": "Lifter Gloss, Hydrating Lip Gloss with Hyaluronic Acid, High Shine for Plumper Looking Lips, Maybelline New York",
  "product_rating": 4.6,
  "review_count": "332",
  "canonical_url": "https://shop.tiktok.com/us/pdp/lifter-gloss-by-maybelline-new-york-hydrating-lip-gloss-with-hyaluronic-acid/1729447567685489298"
}

These are trimmed, live-verified rows; product_price_info, seller_info, review_overviews, and product_card each carry more fields than shown here.

A product that cannot be read

Any product whose request fails is logged with a warning and skipped, and the run continues with the remaining product_ids rather than ending. This holds whether the ID is unresolvable or the request failed for another reason, so one bad entry in a list of ten costs that entry alone. A run in which every product fails still finishes normally with zero rows saved, and the warnings in the log are where the reason is recorded. The one input rejected before any request is made is an empty product_ids list.

Whether maxItems drives further requests

It does not. None of the three carousels paginate. you may like, more from this shop, and the top-reviewed list are each returned complete in a single response, however long TikTok made that response, so maxItems only trims what came back. A product whose you may like rail carries six cards returns at most six rows for that product no matter how high maxItems is set.

What source_product_id points at

source_product_id always names the product being read, never the product being recommended. Running the Actor against several product_ids with scraperType set to topReviewed produces one source_product_id per input product, each attached to that seller's own best-reviewed cards. The rating carried on those cards belongs to the recommended products, not to the input product.

Related Actors

For a single product's own full record (description, images, reviews, and category path), use the TikTok Shop Product Scraper. For discovering products by keyword rather than starting from a product page, use the TikTok Shop Search Scraper. For what surrounds a keyword search instead of a product page (related terms, the shops ranking for it, or its own recommendation widgets), use the TikTok Shop Search Insights Scraper. For what TikTok Shop is featuring on its homepage rather than on one product's page, use the TikTok Shop Trending Scraper.