silkfred.com — AI Site Grade

Analysis

SilkFred — AI-Visibility Audit

The site is a read-only ghost of a defunct business: SilkFred entered administration on 29 October 2025, yet its robots.txt aggressively blocks every major AI crawler while the site continues serving full product pages with "Add to Cart" pricing and "New In" collections — a contradiction that creates a significant AI-knowledge gap.

Crawler Access

The robots.txt explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, Google-Extended, PerplexityBot, Applebot-Extended, and Bytespider — every major AI training and answer-engine crawler — with Disallow: /. However, compare_bot_access shows that all these bots receive a 200 with full HTML content (96 KB, identical to browser baseline). The server (nginx on Ubuntu behind CloudFront) does not enforce the robots.txt directives at the HTTP level. The llms.txt returns a 404. The sitemap.xml is a sprawling index of 30+ sub-sitemaps covering ~19,000 product URLs, yet the /faqs page and /selling page (linked from the footer) both return 404s.

Cold-Knowledge Gap

The LLM prior describes SilkFred as a thriving UK fashion marketplace founded in 2015, acquired by THG in 2021, with Trustpilot reviews and active influencer marketing. The actual site tells a different story: the company entered administration in October 2025, the "About" page uses past tense ("we brought you unique fashion... we did the sourcing"), and the "Sell with Us" page is a 404. The LLM knows nothing about the administration — a critical gap. Any AI engine answering "what is SilkFred?" would describe an active marketplace, not a wound-down entity.

Schema Posture

The homepage carries only Organization and WebSite schema — no Product, Offer, BreadcrumbList, FAQPage, or ItemList markup on any page checked. Category pages (/womens/clothing/dresses, /collection/new-in, /outlet) have zero JSON-LD schema at all. The canonical URLs point to http:// (not https://), creating a mixed-content signal. Despite listing dozens of products with prices and sale discounts, no structured product data is exposed to crawlers.

External Signals

External search results for SilkFred's administration return zero indexed results — no press coverage, no Reddit threads, no news articles about the Quantuma administration surfaced. The Quantuma FAQ page itself returns a 403 Access Denied. The only external links on the site are to Instagram, Facebook, Pinterest, and the blocked Quantuma page. The brand exists in a vacuum: no reviews, no press, no community discussion is findable via search, which means AI models have no secondary sources to correct the stale "active marketplace" prior.

Findings

1. Robots.txt blocks all major AI crawlers High

   The robots.txt explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, Google-Extended, PerplexityBot, Applebot-Extended, and Bytespider with Disallow: /, preventing AI crawlers from indexing the site.

   What to change: Remove the AI crawler disallow rules from robots.txt to allow AI crawlers to index the site.

2. llms.txt file returns 404 Medium

   The llms.txt file, which provides AI crawlers with a curated overview of the site, is missing and returns a 404 error.

   What to change: Create an llms.txt file that lists key pages and provides a summary of the site's content.

3. LLM prior describes SilkFred as an active marketplace High

   The LLM prior knowledge describes SilkFred as a thriving UK fashion marketplace, but the company entered administration in October 2025. This creates a critical knowledge gap where AI models present outdated information.

   What to change: Update the site to clearly indicate the administration status and ensure external sources (press releases, news articles) are indexed.

4. Category pages lack any JSON-LD schema Medium

   Category pages such as /womens/clothing/dresses, /collection/new-in, and /outlet have zero JSON-LD schema markup, missing opportunities for rich results.

   What to change: Add JSON-LD schema markup (e.g., ItemList, Product, Offer) to category and product pages.

5. Canonical URLs use http:// instead of https:// Low

   The canonical URLs on the site point to http:// versions, creating a mixed-content signal that may confuse crawlers.

   What to change: Update canonical URLs to use https:// consistently.

6. FAQs and Selling pages return 404 Medium

   The /faqs and /selling pages, linked from the footer, return 404 errors, indicating broken internal links.

   What to change: Remove or redirect the broken links to relevant pages.

7. No external search results for SilkFred administration High

   Web searches for SilkFred's administration return zero indexed results, meaning AI models have no secondary sources to correct the stale prior.

   What to change: Ensure press releases or news articles about the administration are published and indexed by search engines.

8. Quantuma FAQ page returns 403 Access Denied Medium

   The Quantuma FAQ page that likely contains information about SilkFred's administration is blocked with a 403 error, preventing crawlers from accessing it.

   What to change: Ensure the Quantuma FAQ page is publicly accessible and not blocked by access controls.

9. Homepage schema limited to Organization and WebSite Medium

   The homepage only includes Organization and WebSite schema, missing Product, Offer, BreadcrumbList, and other relevant markup.

   What to change: Add Product, Offer, and BreadcrumbList schema to the homepage and product pages.

10. Sitemap includes 19,000 URLs for a defunct site Low

    The sitemap.xml contains 30+ sub-sitemaps with ~19,000 product URLs, but the site is no longer trading, wasting crawl budget.

    What to change: Reduce the sitemap to only include essential pages or remove it entirely.

What's working

• Server returns full HTML content to all bots despite robots.txt — All AI crawlers receive a 200 response with full HTML content (96 KB), meaning the server does not enforce the robots.txt disallow rules at the HTTP level, allowing content to be accessed if directives are ignored.
• Sitemap index is present and accessible — The sitemap.xml is accessible and contains 30+ sub-sitemaps, providing a comprehensive list of URLs for crawlers.
• About page uses past tense indicating closure — The About page uses past tense language ('we brought you unique fashion'), which may help some crawlers infer the site is no longer active.
• Homepage includes Organization and WebSite schema — The homepage has basic JSON-LD schema for Organization and WebSite, providing some structured data to crawlers.