dior.com

find-all-products

Installation

Adds this website's skill for your agents

 

Summary

Enumerate every product in Dior's storefront for a given locale via the public Algolia search index for /fashion (~5.7k products per locale, 18 API calls) and the per-locale beauty sitemap for /beauty (~350 products). Returns objectID, name, category path, price, color, material, image, URL, and stock state per product. Read-only.

FIG. 01
FIG. 02
FIG. 03
FIG. 04
SKILL.md
213 lines

Dior Find All Products

Purpose

Enumerate every product currently offered on Dior's storefront for a given locale and return one record per product (objectID, name, category path, price, color, material, primary image asset, product URL, stock availability). The fashion (couture) catalog is served by a public Algolia search index — full catalog dumps in ~18 HTTP calls, no auth flow, no anti-bot. The beauty catalog runs on a separate Salesforce Commerce Cloud (SFCC/Demandware) stack with no public JSON API; the recommended path there is the per-locale sitemap (/{locale}/beauty/sitemap.xml) plus optional stealth-browser hydration of individual product detail pages. Read-only — never adds to cart, places orders, or signs in.

When to Use

  • Daily or weekly snapshot of the entire fashion catalog for a locale (price + availability monitoring, new-arrival detection, color/material/category analytics).
  • Cross-locale comparison (US vs UK vs Japan vs France assortments and pricing).
  • Building a search index, product feed, or affiliate catalog mirror.
  • Fashion is the primary path (~5.7k products, fully structured JSON). Beauty (~350 products) is sitemap + browser scrape.

Workflow

Recommended method — fashion: public Algolia API (no browser needed)

Every Dior /{locale}/fashion/* listing page is rendered client-side from a public Algolia search index. The Algolia App ID + search-only API Key are inlined in window.__ENV__ of any fashion page and are stable across locales — they are the production keys, not staging. The endpoint requires no cookies, no Referer, no stealth, and no proxy. The only constraint is Algolia's per-query pagination cap (paginationLimitedTo ≈ 1000 hits per query for these keys, browse ACL is disabled), so a full catalog dump slices by category_lvl0 / category_lvl1 facets.

Region → App ID → API Key → index name (verified 2026-05-21 against dior.com/{locale}/fashion):

Region (cloud)App IDSearch API KeyIndex name patternVerified locales
USC1J7AZ51072c1e320cc4d942713a65869b99252740search_prod_live_{locale}en_us (5,736 products)
EU6EGOOSX8175d568623ff9f43194a18399a53d04ae4search_prod_live_{locale}fr_fr (5,946), en_gb (5,971)
JP9CHAZ88O0Kfeb72a58eb3214acf582857bc6e8e04asearch_prod_live_{locale}ja_jp (5,537)

For any other locale, pull window.__ENV__ from one of its fashion pages (e.g. https://www.dior.com/en_int/fashion/womens-fashion/all-ready-to-wear) and read ALGOLIA_CLOUD_{REGION}_APP_ID / ALGOLIA_CLOUD_{REGION}_API_KEY / ALGOLIA_CLOUD_{REGION}_LOCALES to map locale → region. The fashion page HTML is itself protected by Akamai for some network shapes, so use a stealth Browserbase session (browse cloud sessions create --verified --proxies) to fetch it the first time per locale, cache the result, and never re-do the bootstrap fetch.

  1. Discover bucket sizes for the locale:

    POST https://{APPID}-dsn.algolia.net/1/indexes/search_prod_live_{locale}/query
X-Algolia-Application-Id: {APPID}
X-Algolia-API-Key: {KEY}
Content-Type: application/json
Body: { "query": "", "hitsPerPage": 0, "facets": ["category_lvl0","category_lvl1"] }

    The response's facets.category_lvl0 is a small map (5 buckets for en_us: Women=3405, Men=1567, Baby Dior=514, Jewellery=152, Timepieces=96). Anything ≤ 1000 is fetchable in one query; anything > 1000 needs lvl1 drill-down (Women + Men in en_us).

  2. Fetch each bucket. For buckets ≤ 1000:

    Body: {
  "query": "",
  "hitsPerPage": 1000,
  "filters": "category_lvl0:\"Baby Dior\"",
  "attributesToRetrieve": ["objectID","name","subtitle","description","price","minimumPrice","categories","category_lvl0","category_lvl1","category_lvl2","category_lvl3","color.label","color.code","material.label","damAssets.defaultView","variants.size","stock.hasStock","levelOfStocks","availableOnline","isNew","is_bestseller","collection","saison_diorcom"],
  "attributesToHighlight": []
}

    For buckets > 1000 (Women, Men in en_us), first re-facet within the bucket:

    Body: { "query":"", "hitsPerPage":0, "filters":"category_lvl0:\"Women\"", "facets":["category_lvl1"] }

    then iterate category_lvl1 values and send one query per (lvl0, lvl1) pair:

    Body: { "query":"", "hitsPerPage":1000, "filters":"category_lvl0:\"Women\" AND category_lvl1:\"Clothing\"", ... }

    The deepest split needed in practice is two levels — every (lvl0, lvl1) pair observed on en_us is < 1000.

  3. Construct product URL from objectID: https://www.dior.com/{locale}/fashion/products/{objectID} (e.g. KCO531DET_S30T → …/fashion/products/KCO531DET_S30T). The objectID is {styleCode}_{colorCode} — a SKU-variant key, already deduped at the color level.

  4. Construct image URL from damAssets.defaultView.viewCode: https://assets.christiandior.com/is/image/diorprod/{objectID}_{viewCode}-1?wid=800&hei=1000 (Scene7 path; the viewCode is typically E01/E02/E03). The fashion image base path is hard-coded on the Scene7 CDN — not in window.__ENV__.

  5. Verify totality: sum hits across every bucket and compare to the nbHits of an unfiltered hitsPerPage:0 query. Expect a small (< 1%) gap from products that fall outside the lvl0+lvl1 taxonomy entirely (2 / 5,736 missing on en_us 2026-05-21). To recover them, run a final unfiltered hitsPerPage:1000, page:0 query and union the result.

Total cost for en_us full dump: 18 Algolia POSTs, ~5 seconds wall, $0 LLM. Cross-locale: multiply by however many locales you need.

Recommended method — beauty: sitemap + optional stealth-browser hydration

Beauty runs on SFCC/Demandware (site name dior_us, master library BGXS_PRD). No public JSON API — OCAPI (/s/BGXS_PRD/dw/shop/v22_8/...) returns 404 from the public origin; Sites-DiorUS-Site and Sites-dior_us-Site Demandware controller paths (Search-Show, Product-Show) also 404. The beauty catalog must be enumerated from the per-locale beauty sitemap.

  1. Get the product URL list (no browser, no proxy):

    GET https://www.dior.com/{locale}/beauty/sitemap.xml

    Parse <loc>…/beauty/products/{slug}-{Y\d+}.html</loc> entries. For en_us: 347 product URLs out of 555 total entries (rest are editorial pages, fragrance landing pages, etc.).

  2. Extract product ID from URL: the trailing Y\d{7} token is the SFCC master product ID (e.g. Y0998004 for Sauvage Parfum). Image URLs follow https://www.dior.com/dw/image/v2/BGXS_PRD/on/demandware.static/-/Sites-master_dior/default/dw{hash}/{Y_ID}/{Y_ID}_C{COLOR_ID}_E{VIEW}_{R|G}HC.jpg.

  3. Hydrate per-product detail (optional — only if you need title, price, description, in-stock state). The beauty PDP is server-rendered behind Akamai. Use a stealth Browserbase session:

    sid=$(browse cloud sessions create --keep-alive --verified --proxies | node -e "let s='';process.stdin.on('data',c=>s+=c).on('end',()=>process.stdout.write(JSON.parse(s.match(/\{[\s\S]*\}/)[0]).id))")
browse open "https://www.dior.com/en_us/beauty/products/sauvage-parfum-Y0998004.html" --remote --session "$sid"
browse get title --remote --session "$sid"
browse get html body --remote --session "$sid"

    Title comes back clean (e.g. "Sauvage Parfum: Refillable Citrus and Woody Fragrance | Dior US"). Price strings ($199.00) are inline in the HTML; there is no JSON-LD Product block, so regex the rendered DOM rather than waiting for structured data.

  4. Pace and reuse the session: keep the same stealth session warm across requests (one --keep-alive session can fetch hundreds of beauty pages serially). Don't open new sessions per URL — that triggers Akamai re-challenges.

Browser fallback (if Algolia keys ever rotate or get rate-limited)

The fashion side has a fallback at https://www.dior.com/{locale}/fashion/{category}/all-{type} listing pages (e.g. /en_us/fashion/womens-fashion/all-ready-to-wear). These are CSR React pages — browse snapshot returns few refs, so don't drive clicks. Instead, scroll/lazy-load triggers more Algolia calls; you can also extract window.__ENV__ directly from any fashion page HTML to rediscover the Algolia keys if the values in this skill ever change. Stealth + residential proxy (--verified --proxies) is mandatory for direct fetches; a bare proxy fetch lands on an Akamai challenge page (sec-if-cpt-container div, Powered and protected by Akamai).

For beauty there is no API fallback — the sitemap + stealth browser path is the only path.

Site-Specific Gotchas

  • Two different storefronts share dior.com: /fashion/* is a React SPA driven by Algolia; /beauty/* is a Salesforce Commerce Cloud (Sites-dior_us-Site) classic-rendered site. Their product ID spaces don't overlap (fashion = {8-9chars}_{4chars} like KCO531DET_S30T; beauty = Y\d{7} like Y0998004), their image CDNs are different (assets.christiandior.com/is/image/diorprod/ vs www.dior.com/dw/image/v2/BGXS_PRD/on/demandware.static/), and the data-acquisition strategy is different for each. Always handle them as two separate pipelines, then merge.
  • The Algolia search key is search-only: listIndexes returns 403 ("missing ACL list"), the browse endpoint returns 403 ("missing ACL browse"), and paginationLimitedTo caps query results at the first 1000 hits per request — nbPages will be 1 even when nbHits is 5,736. Slice by facet (category_lvl0 then category_lvl1) to get the rest. Don't bother trying page: 1, 2, 3, ... — those return empty.
  • merch_prod_live_en_us is the wrong index for product enumeration: it lives on a different app (KPGNQ6FJI9), holds 6,419 entries, and includes "push" / promotional card objects (objectID prefix push-NNNNNN) interleaved with real products. The clean catalog is search_prod_live_{locale} on the CLOUD_{region} apps.
  • Locales are split across three Algolia apps: ALGOLIA_CLOUD_US_LOCALES=en_us, ALGOLIA_CLOUD_EU_LOCALES=fr_fr,en_gb, ALGOLIA_CLOUD_JP_LOCALES=ja_jp — these are the only locales explicitly mapped in window.__ENV__ as of 2026-05-21. Other dior.com locales (e.g. en_int, es_es, de_de, it_it) likely route to the EU app but were not verified in this skill's iteration. Test before trusting.
  • merch_prod_live_{locale} indexes are 403 from the CLOUD apps and 200 from MERCH, but the MERCH app does not have search indices for other locales — the apps are partitioned by role (search vs merchandising) AND by region. Don't cross-query.
  • The PRODUCTS_FRA index on the CDC app (N0ZEMF08FA) is the internal master catalog: 352,729 entries, French-language metadata, exploded to every size/color variant, includes internal codes (reference_style_couleur, code_distribution, talendUpdatedAt). It's a leaked internal index — useful for deep enrichment if needed, but not the right surface for a customer-facing product list. The customer-facing surface is search_prod_live_{locale}.
  • objectID is style+color, not style alone: a single named product like "Dior Book Tote" appears once per color variant (e.g. M1296ZRIW_M928, M1296ZRIW_M928… — same first 9 chars, different suffix). If you want deduped "design" rows, group by the prefix before _. If you want every purchasable color, keep the objectID as-is.
  • No url / permalink / slug field in the Algolia hit — construct it: https://www.dior.com/{locale}/fashion/products/{objectID}. The same objectID is the canonical path segment used by the React router.
  • Image URL is constructed, not stored: only damAssets.defaultView.viewCode is in the hit. Full URL: https://assets.christiandior.com/is/image/diorprod/{objectID}_{viewCode}-1?wid=800&hei=1000. Scene7 supports wid/hei/scale params for resizing; preset names like r4x5listing are also valid.
  • Sitemap is bigger than Algolia: /{locale}/fashion/sitemap.xml has 6,104 product URLs for en_us, but Algolia reports 5,736 live products. The ~370-product gap is archived/discontinued URLs that still return a 200 on the PDP (with a "no longer available" message) but are excluded from the search index. If you want purchasable products, use Algolia; if you want every URL Dior has ever shipped under /fashion/products/, use the sitemap.
  • Beauty has no JSON-LD: the SFCC PDP renders product info as plain HTML — no <script type="application/ld+json"> Product block, no inline window.__APP_CONFIG__ style global. Title is in <title>, prices in the visible DOM as $NNN.NN strings. Plan for regex-on-DOM, not JSON parsing.
  • Akamai challenges direct PDP fetches: browse cloud fetch --proxies https://www.dior.com/en_us/fashion/products/{objectID} returns a 403 "Page unavailable" page (with Reference ID and an Akamai challenge script). Stealth + verified + residential proxies in a real Chrome session passes. Listing pages have the same protection. Algolia bypasses Akamai entirely — it's a third-party host (*.algolia.net).
  • Don't waste time on: SFCC OCAPI (/s/BGXS_PRD/dw/shop/v22_8/products/* → 404), Demandware Search-Show / Product-Show controllers under Sites-DiorUS-Site or Sites-dior_us-Site (both 404 from the public origin), https://api-fashion.dior.com/* (private, requires JWT from a logged-in dior.com session), the GraphQL endpoint pattern (none observed for fashion or beauty).

Expected Output

{
  "locale": "en_us",
  "fashion_total": 5736,
  "beauty_total": 347,
  "products": [
    {
      "section": "fashion",
      "objectID": "KCO531DET_S30T",
      "style_code": "KCO531DET",
      "color_code": "S30T",
      "name": "Dior Dentelle Slide",
      "subtitle": "Trench beige embroidered cotton with Dior Oblique motif and white lace",
      "description": "Designed by Jonathan Anderson for the House, the Dior Dentelle slide is a subtle new take on Dior codes of elegance...",
      "categories": ["Women", "Shoes", "Sandals"],
      "category_lvl0": "Women",
      "category_lvl1": "Shoes",
      "category_lvl2": "Sandals",
      "price": { "value": 890, "currency": "USD" },
      "color": { "label": "Beige", "code": "S30T" },
      "material": { "label": "Cotton" },
      "variants": [{ "size": "T34" }, { "size": "T345" }, { "size": "T35" }],
      "stock": { "hasStock": true, "level": "normal" },
      "available_online": true,
      "is_new": false,
      "url": "https://www.dior.com/en_us/fashion/products/KCO531DET_S30T",
      "image_url": "https://assets.christiandior.com/is/image/diorprod/KCO531DET_S30T_E02-1?wid=800&hei=1000"
    },
    {
      "section": "beauty",
      "product_id": "Y0998004",
      "slug": "sauvage-parfum",
      "name": "Sauvage Parfum",
      "url": "https://www.dior.com/en_us/beauty/products/sauvage-parfum-Y0998004.html",
      "price": "$199.00",
      "image_url": "https://www.dior.com/dw/image/v2/BGXS_PRD/on/demandware.static/-/Sites-master_dior/default/dw8bef6812/Y0998004/Y0998004_C099600455_E01_RHC.jpg?sw=800",
      "source": "sitemap+pdp"
    }
  ],
  "discovery_meta": {
    "fashion_index": "search_prod_live_en_us",
    "fashion_app": "C1J7AZ5107",
    "fashion_queries_used": 18,
    "fashion_unique_ids": 5734,
    "fashion_reported_nb": 5736,
    "fashion_coverage_pct": 99.96,
    "beauty_sitemap": "https://www.dior.com/en_us/beauty/sitemap.xml",
    "beauty_urls_in_sitemap": 555,
    "beauty_product_urls": 347
  }
}