Blibli Drop-shipper Data Sync

Keep your reseller catalogue accurate. Give it the Blibli SKUs you sell and it returns the current price, availability, stock and description — plus exactly what changed since the last run, so you push updates instead of re-importing everything.

Built for dropshippers reselling Blibli stock on their own store, Shopee, Tokopedia, TikTok Shop or a webshop.

No login, no API key. HTTP-only — no browser.

What it's actually for

Run it on a schedule. Each run tells you, per SKU:

changeType	What you do in your store
PRICE_INCREASED / PRICE_DECREASED	Reprice — changes.price has from/to and deltaPercent
OUT_OF_STOCK	Pause or hide the listing
BACK_IN_STOCK	Re-activate it
NAME_CHANGED	Update your title
BECAME_UNAVAILABLE	The listing is gone — unpublish
NEW	First time seen
UNCHANGED	Nothing to do

Set onlyChangedItems: true and the dataset contains only the rows you need to act on — often a handful out of hundreds.

{
  "sku": "BLL-70058-00122",
  "syncStatus": "OK",
  "name": "SanDisk Ultra microSDXC 64GB C10 UHS-I Card 100MB/s",
  "price": 255000,
  "listPrice": 299000,
  "discountPercent": 15,
  "inStock": true,
  "stockQuantity": 96,
  "changeType": ["PRICE_DECREASED"],
  "changes": { "price": { "from": 299000, "to": 255000, "deltaPercent": -14.72 } },
  "hasChanged": true
}

Input

{
  "productUrls": ["BLL-70058-00122", "COP-60023-00933"],
  "detailedMode": false,
  "trackChanges": true,
  "onlyChangedItems": true
}

productUrls takes bare SKUs or full product URLs — mix freely. You only need to store SKUs; the actor reaches any product from its SKU alone.

Field	Description
detailedMode	Off: bulk lookup, 1 request per 50 SKUs, gives price + in-stock boolean. On: opens each product page for exact stock quantity, description, warranty, categories and all images — 1 request per SKU
verifyMissing	Re-check missing SKUs on their product page before calling them gone. Leave this on — see below
trackChanges	Compare against the previous run and label each SKU
onlyChangedItems	Emit only rows that changed
stateStoreName	Named store holding the snapshot; use one per catalogue
proxyConfiguration	Residential + country ID recommended for scheduled syncs

Why verifyMissing matters

Blibli's bulk endpoint silently omits some products that are perfectly alive. Verified during recon: SKU VAM-43292-00654 returns nothing from the bulk endpoint while its product page shows a buyable listing with stock 6, and search reports it as AVAILABLE.

If the actor treated "missing from bulk" as "delisted", it would tell you to unpublish products that are still selling. So every missing SKU is re-checked against its product page before being reported as NOT_FOUND. If you switch verification off for speed, those rows say explicitly in _errorDetail that the absence was not confirmed.

Suggested setup

• Price/stock sync, 2–4× daily: detailedMode: false, onlyChangedItems: true. A 500-SKU catalogue is ~10 requests.
• Full content refresh, weekly: detailedMode: true, onlyChangedItems: false — pulls descriptions, warranty and stock levels for rewriting listings.

Known limits

• Fast mode gives an in-stock boolean, not a quantity. Exact stock needs detailedMode, at one request per SKU.
• Variants are flagged but not enumerated — hasVariants tells you a listing has options, but price and stock describe the default variant.
• A run is a point-in-time snapshot; freshness is whatever your schedule is.
• Descriptions are in Indonesian, as Blibli publishes them.

Reliability

Blibli sits behind Cloudflare. The actor clears it HTTP-only with TLS impersonation plus the required Referer, retrying with exponential backoff and rotating both fingerprint and (with a proxy) egress IP on a block. A failing SKU never sinks the run — it becomes a row with _error and _errorDetail while the rest continue.

SKUs absent from a given run are carried forward in the snapshot rather than dropped, so syncing part of your catalogue does not erase the rest of its change history.

See CRAWLING_METHOD.md for the endpoint evidence, the field-by-field comparison of the two modes, and the silent-omission finding.