Suning Scraper

Scrape suning.com — one of China's largest electronics e-commerce platforms (Suning 苏宁易购).

Modes

Mode	Description
search (default)	Products by free-text keyword (Chinese queries work best)
byCategory	Browse a verified category listing (17 categories, e.g. 0-20006-0 手机)
byProductUrl	Fetch one product detail page per URL/ID

Filters

minPrice / maxPrice (CNY), containsKeyword (title substring), sortType (5 orderings: 综合/销量/评价数/价格升降), maxItems.

Output fields

productId, title, sellPoint, specs[], shopName, commentCount, price (CNY), arrivalPrice (到手价 / final price after discounts), vendorName, supplierCode, promotions[], imageUrl, url, sourceUrl, scrapedAt, recordType. Typed error records for invalid URLs / fetch failures / parse failures.

Data Source

Search uses Suning's mobile SSR listing (m.suning.com/search/<query>/) — the desktop search.suning.com endpoint hard-rate-limits (HTTP 429) non-mainland IPs, so the actor reads the mobile SSR instead (same product catalog, paginates via ?page=N, sorts via ?st=N). Category listings (list.suning.com) and product detail pages (product.suning.com) are server-side rendered and freely accessible. Prices come from Suning's public ds.suning.com/generalForTile JSONP endpoint — no login, no API key. The actor fetches pages direct first and escalates to the Apify AUTO proxy on 429/403/5xx.

Canonical product URLs: every record's url/sourceUrl uses the canonical form product.suning.com/{pid}.html. Suning's own SSR links use a shop-prefixed form (product.suning.com/{shop}/{pid}.html) that returns 404 from a clean shell — the actor always rewrites to the canonical form, which serves HTTP 200.

Anti-bot

Tier 2: curl_cffi chrome131 TLS impersonation with rotating browser user-agents, exponential backoff on 429/5xx, lazy Apify AUTO-datacenter proxy escalation (direct-first, proxy fallback).

Limits

• Prices come from Suning's public ds.suning.com/generalForTile JSONP endpoint, which serves 404 to some non-mainland egress IPs. The actor retries each price batch across fresh proxy sessions (some AUTO exits are mainland) and attaches prices whenever a mainland exit is found; otherwise records ship without the price fields (titles, sell points, specs, images, URLs are always populated from the SSR listing HTML). Some pre-release / out-of-stock items legitimately carry no price on either the SSR or the price API.
• minPrice/maxPrice are strict: when a price bound is set, records without an attached price are excluded (the filter always reflects in the output; an impossible range like minPrice > maxPrice yields 0 records cleanly).
• promotions may be empty when no promotion is active.
• Mobile search ignores st=10 (price descending) and falls back to the default ordering — the field still validates and runs.
• Mobile SSR pages expose 10-20 products per page; maxItems above a single page's yield walks ?page=N until filled.