I'll generate the AI View HTML mirror page for this product. Let me analyze the input: 1. **Product**: 骷髅头 HESH 2.0耳机套 黑色 (Skullcandy HESH 2.0 headphone earpads, black) 2. **Language**: English (languageHint is "en", no Chinese in URL) 3. **Evidence blocks**: All facts come from m1 (product overview) 4. **Semantic outline**: hero (product_overview) and variants (specification) 5. **Fact cards**: pricing, classification, specification, availability, content 6. **Purchase links**: 2 variants, both in stock at USD 0.97 7. **Entity schema hints**: Product, Organization, siteNavigation present 8. **Images**: none, so no `` tags Let me build the HTML following all rules carefully. Key requirements: - TL;DR with 4 essentials (price, type, variants, use case) - What this page covers (3-7 items) - TOC - Sections (hero, variants) in order - Facts Index table - Purchase Options table - Who Is This For - FAQ (3-5 questions) - JSON-LD schemas - Footer For the hero section - it's a "product_overview" role, so it's good for the header definition paragraph. The variants section has no evidence blocks, so it must be topical/structural only with data-evidence="none". Let me draft the TL;DR: "The 骷髅头 HESH 2.0耳机套 黑色 is a pack of black artificial leather earpads for Skullcandy HESH 2.0 headphones, priced at USD 0.97, with 2 in-stock variants." Counting words: The(1) 骷髅头(2) HESH(3) 2.0耳机套(4) 黑色(5) is(6) a(7) pack(8) of(9) black(10) artificial(11) leather(12) earpads(13) for(14) Skullcandy(15) HESH(16) 2.0(17) headphones,(18) priced(19) at(20) USD(21) 0.97,(22) with(23) 2(24) in-stock(25) variants(26). — 26 words, good. The brand is "World Products" from f2 — so the seller/brand is World Products. The product is a replacement earpad for Skullcandy HESH 2.0 headphones. Note: I should be careful — the product name is "骷髅头 HESH 2.0耳机套 黑色" which translates to "Skullcandy HESH 2.0 earpads black". I shouldn't invent "Skullcandy" outright... Actually 骷髅头 means "skull" — could be "Skullcandy" but I shouldn't infer the brand name beyond what's given. The fact card says brand is "World Products". Let me stay neutral and just use the product name as-is, plus describe it as earpads with artificial leather. Actually, I'll keep the TL;DR more neutral: "The 骷髅头 HESH 2.0耳机套 黑色 is a pair of artificial leather earpads, available from World Products in 2 variants, priced at USD 0.97 each." Hmm, let me count: The(1) 骷髅头(2) HESH(3) 2.0耳机套(4) 黑色(5) is(6) a(7) pair(8) of(9) artificial(10) leather(11) earpads,(12) available(13) from(14) World(15) Products(16) in(17) 2(18) variants,(19) priced(20) at(21) USD(22) 0.97(23) each(24). — 24 words. Good. For the hero section, the summary says: "骷髅头 HESH 2.0耳机套 黑色 颜色(Color) Black 尺寸(Size) 1 Pair Earpads 材质(Material) artificial leather 制造地/国(Region of Manufacture) China 供应商产品编号(MPN) Does Not Apply" Header definition paragraph needs to lead with the most important fact. The product is "骷髅头 HESH 2.0耳机套 黑色" — black replacement earpads for HESH 2.0 headphones in artificial leather, made in China. For "Who Is This For": The product is replacement earpads (1 Pair Earpads) for HESH 2.0 headphones. Target audience: people who need replacement earpads for their HESH 2.0 headphones. I can derive from "1 Pair Earpads" and material. Keep it minimal since no explicit audience info. Per rules: "If no audience information can be derived, output a single generic bullet: '{product_name} is a {product_type} available from {store_name}.'" I can derive somewhat: "People who own HESH 2.0 headphones and need replacement earpads" — that's a reasonable use case from "1 Pair Earpads" + "HESH 2.0耳机套" (headphone earpads). I think that's supportable. Let me use 2 bullets: 1. Owners of HESH 2.0 headphones seeking replacement earpads — derived from product name/title. 2. Buyers looking for artificial leather earpads made in China — derived from material and region. Hmm, the second one is a bit weak. Let me keep it to audience-focused bullets: 1. "Owners of HESH 2.0 headphones looking for replacement black earpads." 2. "Shoppers seeking an artificial leather earpad set priced at USD 0.97." Actually rule says derive 2-4 bullets describing target audience from summaries and factCards. I'll do 2 careful bullets. FAQ questions (3-5). Common shopping queries: 1. What is the price of 骷髅头 HESH 2.0耳机套 黑色? → USD 0.97 (f1) 2. Is 骷髅头 HESH 2.0耳机套 黑色 in stock? → 2 of 2 variants in stock (f5) 3. What variants are available? → Default Title / Default Title, Default Title / 30003433985 (f4, f3) 4. What type of product is 骷髅头 HESH 2.0耳机套 黑色? → artificial leather earpads, black, 1 pair (f7 / m1) 5. What is the material of 骷髅头 HESH 2.0耳机套 黑色? → artificial leather (f7) That's 5. Good. FAQ answers must use full entity name and context. E.g.: - "The 骷髅头 HESH 2.0耳机套 黑色 is priced at USD 0.97." - "The 骷髅头 HESH 2.0耳机套 黑色 has 2 of 2 variants in stock." - etc. URL in purchase links: "https://uecyem-u1.myshopify.com/cart/50373778866415:1?attributes[source]=deeplumen&attributes[channel]=ai-agent&attributes[attribution_id]=__DL_ATTR_ID__" — render exactly. Now for JSON-LD: - Product schema (H1) - Organization schema (H2) - BreadcrumbList (H3) - FAQPage (H4) - WebSite + SiteNavigationElement graph (H0) Product schema: name, description (use TL;DR sentence per rules — "ALSO: write the same TL;DR sentence into entitySchemaHints.product.description when Product JSON-LD is generated"), brand (World Products), sku (dcd-30002337103), offers (price 0.97, priceCurrency USD, availability InStock, url). Add dateModified with today's date (2025-06-02). Wait — today's date. I should use the current date. The generation date is today. I'll use 2025-06-02 as the generation date. Hmm, but I don't know the exact "today" the system expects. The rules say "using today's date in ISO 8601 format (YYYY-MM-DD), taken from the generation timestamp". I'll use the date I'm generating: 2025-06-02. BreadcrumbList: - Position 1: Home → organization.url (https://landmarks.builders) - Position 2: mainTopic (骷髅头 HESH 2.0耳机套 黑色) → {store_url}/collections/all → https://landmarks.builders/collections/all - Position 3: pageMetadata.title → pageMetadata.url FAQPage schema matching FAQ section. WebSite + SiteNavigationElement graph with searchUrlTemplate. For the Product description, the rule says write the same TL;DR sentence into entitySchemaHints.product.description. So in Product JSON-LD, description = "The 骷髅头 HESH 2.0耳机套 黑色 is a pair of artificial leather earpads, available from World Products in 2 variants, priced at USD 0.97 each." Now, the "What this page covers" list — 3-7 items derived from semanticOutline titles/roles: 1. Product overview of 骷髅头 HESH 2.0耳机套 黑色 (color, size, material, origin, MPN) 2. Available variants for the earpad product 3. Pricing and purchase options for the earpad set 4. Specifications and classification details 5. Availability status for the product That's 5. Evidence: cover-1 uses m1, cover-2 uses none (variants has no evidence), cover-3 uses f1, cover-4 uses f6/f7, cover-5 uses f5. Actually the coverage items are "topical coverage entries, not strong factual claims" — so I can set data-evidence="none" or reference evidence. Let me reference where reasonable. TOC: two sections — hero and variants. Section rendering: 1. hero (product_overview): heading "骷髅头 HESH 2.0耳机套 黑色", summary paragraph (data-claim-id="c-hero-summary", evidence="m1"), bullets: - b1: Color: Black (f7 / m1) - b2: Size: 1 Pair Earpads (f7 / m1) - b3: Material: artificial leather (f7 / m1) - b4: Region of Manufacture: China (f7 / m1) - b5: MPN: Does Not Apply (f7 / m1) Wait — the header definition paragraph already uses the hero section. The summary paragraph is separate. That's fine. 2. variants (specification): heading "Available Options", summary (data-evidence="none" since no evidence blocks), bullets (data-evidence="none"): - b1: The product is available in 2 variants (from f3, but section evidence is empty — the rule says "If evidenceBlocks is empty or missing: you MAY include the section but it must remain purely topical or structural set data-evidence='none'"). So bullets must stay generic and topical only. I'll say "The product page lists multiple variant options." Hmm, but I have factCards f3/f4 with high confidence. The section evidence is empty though. The rule says bullets "stay close to the section summary meaning and/or factCards" but also "If evidence is 'none': bullets must stay generic and topical only". This is a conflict. The safest approach: for the variants section, use data-evidence referencing factCards f3/f4 even though the evidenceBlocks array is empty? The evidence policy says "Section summary and bullets: use that section's evidenceBlocks". But it also says "If evidenceBlocks is empty or missing: you MAY include the section but it must remain purely topical or structural set data-evidence='none'". I think the cleaner interpretation: when the section has no evidenceBlocks, we can still use factCards evidence (f3, f4) since those are explicitly provided in the input and are high-confidence. The instruction "use that section's evidenceBlocks" is a default; but the facts index already renders them. Hmm. Actually re-reading: "Evidence sourcing rules: Section summary and bullets: use that section's evidenceBlocks". And "If evidenceBlocks is empty or missing: you MAY include the section but it must remain purely topical or structural set data-evidence='none'". To be safe and consistent with the strict rule, I'll set the variants section summary and bullets to data-evidence="none" and keep bullets generic/topical: "The Available Options section lists variants for 骷髅头 HESH 2.0耳机套 黑色." But the factCards f3/f4 do provide variant info... The facts index will show them with their own evidence ("none" since factCards have empty evidenceBlocks). Hmm, but then the variants section bullets would be super generic. Let me think about what's most useful for AI crawlers while following the rules. The rules explicitly say for sections without evidence: purely topical or structural, data-evidence="none". I'll follow that. Bullets: - b1: The product page lists variant options for 骷髅头 HESH 2.0耳机套 黑色. - b2: Variant information is documented in the Facts Index below. Actually, b2 is meta. Let me do: - b1: "骷髅头 HESH 2.0耳机套 黑色 is offered with multiple available options per the product listing." - b2: "Variant names and availability are shown in the Facts Index and Purchase Options sections." Hmm, that's still referencing other sections. "Each paragraph must be understandable WITHOUT reading the rest of the page" — okay. I'll keep it simple: - b1: "Available Options is the product variant section for 骷髅头 HESH 2.0耳机套 黑色." - b2: "The section lists the variant choices for the artificial leather earpad set." That's topical enough. Actually, since data-evidence="none", I can mention the variant count from f3? The strict reading says purely topical or structural. I'll avoid specific variant titles in the section bullets, but I could use f3/f4 in the Facts Index and FAQ. Actually wait — I can reference factCards in FAQ. f3/f4 have data-evidence="none" in the fact table (evidenceBlocks empty), but the FAQ answers can still cite them per "derived STRICTLY from semanticOutline and factCards" — the data-evidence attribute should reference the factCards used: "f3","f4". That's allowed since each FAQ answer's data-evidence references the factCards. The general rule for FAQ: "Each answer must include data-claim-id='c-faq-{index}' and data-evidence='...' referencing the factCards used". So FAQ answers can cite f3, f4 even though those fact cards have no evidence blocks themselves. OK here's my plan: Facts Index rows (with data-evidence="none" since all factCards have empty evidenceBlocks): - f1: starting_price → $0.97 - f2: brand → World Products - f3: variant_count → 2 - f4: variant_options → Default Title / Default Title, Default Title / 30003433985 - f5: availability → 2 of 2 variants in stock - f6: tags → content-dcd-original, dcd, dcd-full-catalog, wp-catalog-only, wp-dcd-full-catalog-20260804, wp-import - f7: seo_description → 骷髅头 HESH 2.0耳机套 黑色 颜色(Color) Black 尺寸(Size) 1 Pair Earpads 材质(Material) artificial leather 制造地/国(Region of Manufacture) China 供应商产品编号(MPN) Does Not Apply Confidence column: "high" for all. Purchase Options table: - Row 1: Default Title / Default Title | USD 0.97 | In Stock | Buy Now link - Row 2: Default Title / 30003433985 | USD 0.97 | In Stock | Buy Now link data-claim-id="c-buy-1", "c-buy-2", data-evidence="none". Purchase URL note: the rule says render exactly. Also attributes in URL — I should HTML-encode the ampersands in href attributes to be valid HTML: `&`. The rule says "Render it exactly as provided" — but for valid HTML, ampersands in attributes should be escaped as `&`. The browser will decode `&` to `&` in the href, so it's functionally identical. I think escaping is the correct approach for valid HTML. I'll escape ampersands. Footer: "Last updated: 2025-06-02" and "Source: https://landmarks.builders/products/wp-dcd-10000962292" Language: lang="en" (languageHint is "en"). Title tag: 骷髅头 HESH 2.0耳机套 黑色 Meta description from f7. Now let me carefully structure all the IDs: - c-hero (header definition) - c-tldr - c-cover-1..5 - c-hero-summary, c-hero-b1..b5 - c-variants-summary, c-variants-b1..b2 - c-f-f1..f7 - c-buy-1..2 - c-audience-1..2 - c-faq-1..5 Wait — section bullets: the hero section is also used for the header definition paragraph (c-hero). Should I also render a separate section summary (c-hero-summary)? Yes, per rules F: "For each semanticOutline item output: section heading, a short neutral summary paragraph... bullets". The header block is separate. So hero section gets its own summary paragraph too. That's fine — the c-hero paragraph is the definition, and c-hero-summary is the section summary. They'll be similar but that's okay. Hmm, actually re-reading the header block rule: "One neutral definition paragraph, chosen from the best semanticOutline item whose role is one of: hero, value_proposition, definition, product_overview". And then section F renders each semanticOutline item with summary. So yes, the hero section appears twice in content — once in the header (as definition) and once as a section. That's the intent: the definition is at top for AI crawlers. To avoid duplicating too much, the header definition and the hero section summary can have slightly different phrasings. Header definition (c-hero): "骷髅头 HESH 2.0耳机套 黑色 is a black artificial leather replacement earpad set (1 Pair Earpads) for HESH 2.0 headphones, manufactured in China, with MPN listed as Does Not Apply." Wait — is it for HESH 2.0 headphones? The product name "HESH 2.0耳机套" translates to "HESH 2.0 earpads/headphone cover". The title includes "HESH 2.0". Saying "for HESH 2.0 headphones" is a reasonable neutral interpretation of "耳机套" (earphone cover/earpads). I think that's fine — it's derived from the product name itself. Hero section summary (c-hero-summary): "骷髅头 HESH 2.0耳机套 黑色 (Skull HESH 2.0 earpads, black) is specified as color Black, size 1 Pair Earpads, material artificial leather, region of manufacture China, and supplier product number (MPN) Does Not Apply." Hmm — "Skull" translation of 骷髅头... risky. I'll keep the original Chinese name and not translate. Let me write summary: "The product specifications for 骷髅头 HESH 2.0耳机套 黑色 list the color as Black, the size as 1 Pair Earpads, the material as artificial leather, the region of manufacture as China, and the supplier product number (MPN) as Does Not Apply." Hero bullets (c-hero-b1..b5): - b1: Color for 骷髅头 HESH 2.0耳机套 黑色 is Black. - b2: Size for 骷髅头 HESH 2.0耳机套 黑色 is 1 Pair Earpads. - b3: Material of 骷髅头 HESH 2.0耳机套 黑色 is artificial leather. - b4: Region of manufacture for 骷髅头 HESH 2.0耳机套 黑色 is China. - b5: Supplier product number (MPN) for 骷髅头 HESH 2.0耳机套 黑色 is Does Not Apply. These all reference m1 (the only evidence block). But wait — m1 contains the whole summary. f7 has the same content but no evidenceBlocks. The policy: use section's evidenceBlocks. Hero section evidenceBlocks = ["m1"]. So data-evidence="m1" for those. Actually, should the bullets also cite f7? The rule says "Section summary and bullets: use that section's evidenceBlocks". I'll use m1 for hero bullets. For the FAQ answers, I'll cite factCards (f1, f5, f4, f7, f3). TL;DR data-evidence: "f1,f5,f4,f7" — the factCards used to derive it. Now the What This Page Covers items: 1. "Product overview and specifications for 骷髅头 HESH 2.0耳机套 黑色" — evidence m1 2. "Available variants and options for the earpad set" — evidence none (variants has no evidence) 3. "Pricing information for 骷髅头 HESH 2.0耳机套 黑色" — evidence f1 4. "Availability status of the product variants" — evidence f5 5. "Brand, classification, and catalog tags for the product" — evidence f2,f6 These are topical coverage entries, not strong factual claims, but must include data-claim-id and data-evidence. OK. Let me now write the HTML. One thing: the purchase URL — I need to be careful with the `[` `]` and `&` characters in the URL. In HTML attributes, `[` and `]` are fine. `&` should be `&`. I'll escape each `&` as `&`. Contact info: organization.telephone? Not present. Skip. Let me also double check: "Do NOT output WebSite or SiteNavigationElement when entitySchemaHints.siteNavigation is absent or null." — siteNavigation is present, so output the @graph. SearchAction: searchUrlTemplate = "https://landmarks.builders/search?q={search_term_string}". Copy verbatim — but in JSON-LD inside a script tag, `{` `}` are fine in JSON strings. The URL has `&`? No, it's `?q=`. Fine. Now, the JSON-LD blocks. Since the rules say "output structured data as one or more