🛍️ Applies to Product Detail only. Controls which upstream detail version is used. They differ in data freshness and pricing depth — pick deliberately:
- 🆕 Fresh — Full — live, non-cached scrape, flat response, stamped with a true-capture
recordTime. ⚠️ Its upstream collector is currently failing almost every request, so a run asking for this version will usually be served by Fresh — Standard instead (see Automatic fallback below) — which returns different field names. Pick Standard directly unless you specifically need this shape.
- 🔋 Fresh — Standard (default, recommended) — live, non-cached scrape with the full classic payload (SKUs, pricing tiers, item description HTML, coupons, shipping, return policy). Also stamped with
recordTime. Coverage differs slightly from Full — some products only resolve on one version, so for maximum coverage you can run both and merge.
- 💎 Premium — final / coupon pricing — live scrape focused on the final, coupon-adjusted price of the item and every SKU variant (
finalPriceYuan, finalSkuPrice, original-vs-final per SKU, stock). Best when accurate checkout pricing is the priority.
- ⚡ Cached — Lite (NOT recommended for fresh data) — compact response served from a cache that can be anywhere from a few days to several months old. Fields may refresh independently and there is no reliable capture timestamp (the
recordTime on a cached row reflects request time, not when the product state was captured). It is the most compact (lightest) payload — only choose it when stale data is acceptable. Note it is billed at the same per-row rate as the standard fresh versions (Full / Standard), so it is not cheaper, just smaller.
🔄 Automatic fallback: the three standard versions (Full / Standard / Lite) cover different products, so if the one you pick cannot return an item the actor automatically tries the others and returns whichever answers. The version that answers decides the field names — they differ a lot between versions (only itemId, title and images are common to all), so read the field matrix in the README before building a pipeline on a specific field. The 💎 Premium version is never substituted in or out: ask for it and you get it, or you get an error row.
💡 Freshness rule: for current-state data use Fresh — Standard (default) or Fresh — Full; for accurate final pricing use the 💎 Premium version; avoid Lite when freshness matters. Check the recordTime field on each row to verify capture time.
