Blibli Homepage Recommendation & Ad Placement Tracker

See what Blibli's own algorithm is placing on its homepage right now — not what a search or category browse returns, but the actual "Untukmu" (For You), "Zona affiliate", and auto-generated category tabs a real visitor sees, with rank order and paid ad slots included.

Built for brand/seller homepage visibility monitoring ("is my product being surfaced, and where"), affiliate product discovery (the "Zona affiliate" tab is a curated affiliate-marketing feed), and homepage-ad competitive intelligence (which brands are buying banner placement, and when their campaign runs).

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

One event, no mode

Every other actor in this developer's Blibli portfolio uses a mode input when a target needs genuinely different endpoints with different output shapes (see blibli-promo-extractor's three modes). This one doesn't, deliberately: every tab hits the same endpoint and returns the same tile shapes. Tabs are a filter dimension of one event — "what's on the homepage feed right now" — not separate events, so they're a tabs array input instead of a mode enum.

{ "tabs": ["TAB_1", "TAB_4"], "maxPagesPerTab": 15 }

Leave tabs empty to auto-discover and sweep every tab Blibli currently exposes (8 at the time of writing — this list itself changes as Blibli adds seasonal tabs, so the actor always re-fetches it rather than hardcoding names).

Output — three tile types in one feed

Blibli's feed interleaves three unrelated tile shapes; _recordType tells them apart. Blibli's own fields are passed through verbatim; fields this actor adds are prefixed with _.

product — a real recommendation:

{
  "_recordType": "product",
  "_tabName": "TAB_1",
  "_tabTitle": "Untukmu",
  "_rankInFeed": 2,
  "name": "Hanasui Collagen Water Sunscreen SPF30 30gr",
  "sku": "HAO-70179-00023",
  "merchantName": "Hanasui Flagship Store",
  "_offerPrice": 30600,
  "_discountPercent": 15,
  "_productUrl": "https://www.blibli.com/p/...",
  "_personalizationSource": "brs",
  "_personalizationLogic": ";;DEFAULT_RERANK"
}

ad_banner — a paid brand placement, no product at all:

{
  "_recordType": "ad_banner",
  "_tabName": "TAB_1",
  "_brandAdId": "BA00002377_homePage",
  "_adBrandNames": ["KGA - New Launch"],
  "_adRedirectUrl": "https://www.blibli.com/promosi/kga-advan-laptop",
  "_adStartTime": "2026-08-27T17:00:00Z",
  "_adEndTime": "2026-09-09T16:59:59Z"
}

category_card — a curated "trending now" collection tile, not a single SKU:

{
  "_recordType": "category_card",
  "name": "Sepatu Sneakers Pria",
  "_cardLabel": "Lagi laris",
  "_cardDescription": "1,7 rb terjual bulan ini",
  "_productUrl": "https://www.blibli.com/top-selling-page/..."
}

_personalizationSource / _personalizationLogic / _strategyId are Blibli's own algorithm-internals fields, passed through — a rare public window into which recommendation strategy served a given placement.

Input reference

Field	Description
tabs	Tab short-codes to sweep (e.g. ["TAB_1","TAB_4"]). Empty = every tab Blibli currently exposes.
maxPagesPerTab	Hard cap on pages per tab (default 15, ~20 items/page). This is a personalized, re-ranked feed with no reliable "end" — see Known limits.
maxItemsPerTab	Stop emitting for a tab after this many records. 0 = no limit beyond maxPagesPerTab.
includeAdBanners	Keep paid ad-banner tiles (default true). Turn off for product recommendations only.
proxyConfiguration	Apify Residential + country ID recommended for long runs.

Known limits

• No absolute "all recommendations." This is a live, re-ranked feed (Blibli's own field literally says

  personalizationLogic: "DEFAULT_RERANK"
  ), not a fixed catalogue — maxPagesPerTab is what actually bounds a run for personalized tabs, not an end-of-list signal. Auto-generated category tabs (AUTOMATIC source) do terminate cleanly on an empty page; personalized tabs (MANUAL source, e.g. "Untukmu") may not, by design of what a recommendation feed is.
• Blibli's own paging.total_page is unreliable on this endpoint — observed oscillating between requests to the same tab. This actor never uses it as a stop condition; see CRAWLING_METHOD.md for the measurement.
• Anonymous/default ranking only. No session or cookies are sent, so this reflects Blibli's no-login ranking, not any specific shopper's personalized feed.

Pricing

$1.00 per 1,000 results — pay-per-event, one event (apify-default-dataset-item), charged per record in the dataset. No subscription, no minimum.

Store category: ECOMMERCE — the actor's output is marketplace product/placement data.

Reliability

Blibli sits behind Cloudflare. The actor clears it HTTP-only with TLS impersonation plus the Referer header Cloudflare's rule requires — no browser, no CAPTCHA solver. On a block it retries with exponential backoff while rotating both the TLS fingerprint and (when a proxy is configured) the egress IP.

Runs never fail silently: a blocked run, an upstream schema change, bad input, or a genuinely empty result each emit a diagnostic row carrying _error and _errorDetail.

See CRAWLING_METHOD.md for the full WAF analysis, endpoint discovery (including the id-vs-name trap and the tile shapes found during local testing), and why paging.total_page can't be trusted.