SEO Audit Case Study: Why Mulan.pro Has Zero SEO Awareness

SEOCase StudyTechnical SEO

Good SEO isn't about how many keywords you stuff into your pages. It's about whether your team built the site with SEO fundamentals in mind from day one.

I recently ran a full SEO diagnosis on Mulan.pro — an AI video workflow product — using 哥飞 SEO Agent. The conclusion was direct: the team has zero SEO awareness.

It's not that a few Meta tags are missing. From the rendering strategy and language routing to robots.txt, sitemap, page tags, and internal site structure — almost none of the basic SEO infrastructure is in place.

Problem 1: Pure Client-Side Rendering Across All Pages

This is Mulan.pro's most critical issue.

When the agent crawled the homepage, /en/, /discover, user pages, and flow pages, the first response for every URL was the same ~2.4KB HTML:

<div id="root"></div>

Zero words of body text in the raw HTML. No H1. No internal links. The homepage, English page, and Discover page all share the same Title: MulanAI, and the same one-word Description: Mulan.

Googlebot gets nothing of substance on its first crawl. It has to execute JavaScript just to see any page content at all.

The fix: A site doing SEO cannot hide all content behind JavaScript. Titles, body text, H1s, internal links, product descriptions, and FAQs must appear in the HTML on the first request.

Start by converting the homepage and key feature pages to SSR, SSG, or pre-rendering. Give each page its own Title, Description, and H1.

Problem 2: Root Directory Is Not a Stable Homepage

Opening https://mulan.pro/ redirects based on the browser's language setting — either /zh or /en.

The root directory / carries no stable default page. Search engines and users hitting the same address see different target URLs depending on their browser language setting.

The better approach: Pick one language as the primary (e.g., English). The root directory serves the default English page. /en should 301 redirect to the root. All other language variants (/zh, /ja, etc.) become independently crawlable pages, each with their own canonical and hreflang declarations. The root serves as the default entry, with x-default indicating the generic入口 to Google.

Problem 3: robots.txt and sitemap.xml Redirect to Homepage

Mulan.pro doesn't serve either of these SEO fundamentals correctly:

  • https://mulan.pro/robots.txt
  • https://mulan.pro/sitemap.xml

Both URLs redirect to the homepage and return the homepage HTML — not robots text or sitemap XML.

robots.txt tells search engines what they may and may not crawl. sitemap.xml is a proactively submitted page inventory. Put the actual files at their correct paths.

Problem 4: Non-Existent URLs Also Redirect to Homepage with 200

Visiting a clearly non-existent path like /this-page-should-not-exist-xyz123 also redirects to the homepage — with a 200 status code.

Non-existent pages should return 404 or 410. Redirecting all unknown paths to the homepage while returning 200 tells search engines "this page exists normally" — a textbook soft 404.

A soft 404: the server says the page is fine, but when Google actually visits it, the page doesn't exist.

Problem 5: Test Environments Got Indexed by Google

The agent found these subdomains in site:mulan.pro results:

  • test.mulan.pro
  • dev-ent.mulan.pro
  • dev-api.mulan.pro

Opening test.mulan.pro's source revealed PUBLIC_MODE: "TEST", testgateway.mulan.pro, and test storage addresses.

Test environments are publicly accessible AND indexed by Google. Users searching for the brand could land on non-production pages.

The fix: Add authentication to test/dev subdomains if they're for internal use only. Add noindex to any publicly accessible preview pages.

Problem 6: No canonical or hreflang

Mulan.pro's raw HTML contains neither canonical nor hreflang tags.

  • canonical tells Google which URL is the canonical (standard) version
  • hreflang declares the language relationships between Chinese, English, and other language pages

The root directory redirects to /zh or /en based on browser language — this makes proper language relationship declarations even more critical. Without them, Google has to guess.

Problem 7: No H1, Internal Links, OG Tags, or Structured Data

Continuing the audit, the agent identified four more gaps:

  • No H1-H6 in the raw HTML
  • No crawlable internal links
  • No Open Graph or Twitter Card meta tags
  • No JSON-LD structured data (SoftwareApplication, Organization, etc.)

H1 is the page's main heading. Internal links connect different pages. OG and Twitter Card determine whether shared links render with a rich card. JSON-LD helps search engines understand what product this is and who provides it.

All of these belong in the HTML returned by the server — not rendered by JavaScript after the fact.

Problem 8: Keywords Concentrated on Brand Terms

SimilarWeb data showed Mulan.pro's estimated July 2026 traffic at 3,203 visits, with the prior two months at 13,460 and 9,524.

All major search terms were brand-related: mulan ai, 木兰 AI, 木兰 AI 融资, mulan pro.

This means everyone finding the site through search already knew the brand. Users searching for AI video tools without knowing Mulan's name have no non-brand path into the site.

Mulan already has video generation, editing, batch production, and workflow capabilities. These could be split into dedicated feature pages targeting text to video, ai video editor, tutorials, FAQs, and specific use cases.

Keyword coverage shouldn't grow indiscriminately. First assess the actual search demand, then decide which page type handles which keyword.

Recommended Fix Priority

If this were my site, I'd fix in this order:

Step 1: Fix client-side rendering and language routing. Convert the homepage and key feature pages to SSR/SSG/pre-rendering so titles, body text, H1, internal links, and FAQs appear in the raw HTML on first request. Stabilize the root directory so /zh and /en become explicit, crawlable language pages.

Step 2: Fix foundational files and error states. Serve real robots.txt and sitemap.xml at their correct paths. Return 404 or 410 for non-existent URLs instead of redirecting everything to the homepage.

Step 3: Isolate test environments. Add authentication to internal test/dev subdomains. Add noindex to any publicly accessible preview pages.

Step 4: Build out page signals and keyword entry points. Add canonical, hreflang, OG tags, structured data, and crawlable internal links. Then split existing product capabilities across feature pages, tutorial pages, and FAQs.

For an already-launched site: don't rush to add new pages. Get existing pages and foundational configuration right first, then decide what content to add.

Summary

Mulan.pro's problems aren't about one missing Title tag — the team simply never built SEO into the development process from the start.

Client-side rendering means Googlebot gets no content on first crawl. The root directory depends on browser language redirects. robots.txt, sitemap, and error URLs all redirect to the homepage. Test environments are indexed. Pages lack canonical, hreflang, H1, internal links, and structured data.

These aren't hard to fix. But the order matters: first make it possible for Google to get the right page directly, then handle crawling, indexing, and page signals, and only then focus on keywords and content.

References

Back to SEO Guide