Hreflang and canonical audit: 192 URLs across 16 languages

The result, without the hype

Every one of the 192 successfully fetched pages returned HTTP 200, carried its own canonical URL, declared the expected HTML language and included reciprocal hreflang alternates across the inspected locale group. Each had 17 raw HTML alternates: 16 language codes plus x-default.

We found zero defects under these specific automated checks. That is a useful release-quality signal for this codebase. It does not prove flawless translations, correct rankings, Google indexing, full-site health, or a general industry defect rate.

Five checks worth automating before an international launch

Exactly which URLs were checked?

For each of the 12 routes below, we fetched the English original and 15 localized counterparts. The raw CSV records each requested URL, final response, canonical, HTML language, alternate count, and issue codes.

Title distinctness is based on exact raw HTML title strings, not on linguistic review. The response and canonical counts are from the same snapshot—not independently verified by Google.

Methodology and replication

We intentionally selected 12 reusable page families: homepage, technical audit, free-tools hub, keyword research, most-visited websites, SEO guide, GSC MCP, sitemap checker, redirect checker, slug generator, HTTP status checker, and internal-link checker. We checked the same URLs in all 16 published languages, for 192 requests total.

1. Fetch each public page over HTTPS, with redirects followed, six concurrent requests, and a 20-second per-request timeout.
2. Read raw server-response HTML only. Extract rel="canonical", rel="alternate" hreflang, html lang, title, description, and robots meta tags.
3. Compare each canonical with its requested URL and every expected locale target with the published route mapping.
4. Verify that reciprocal hreflang references exist in the fetched counterpart pages. Count missing or inconsistent targets separately from network failures.
5. Publish per-URL CSV and the zero-dependency script so another developer can rerun the same checks.

Dataset access and reuse are subject to DigestSEO's website terms; the linked repository's open-source code license does not independently license this research dataset.

The observations come from an automated raw-HTML crawl. Editorial text and translations were AI-assisted; independent native-language review is not claimed.

Limits of this audit

This snapshot excludes other pages and content families, dynamically rendered metadata, Googlebot's rendering, Search Console index status, language quality, crawl budget, server-side personalization, and external backlinks. It is not an industry census or a claim that Google will index all 192 URLs. Results can change after deployment.

What the result means for international SEO teams

A correct hreflang cluster is infrastructure, not a traffic-growth guarantee. Even a technically clean multilingual rollout still needs genuinely localized search intent, useful original content, relevant links, and ongoing search-performance analysis.

Google's documentation requires locale pages to list all versions (including themselves) and maintain return links. Its canonical guidance is separate: canonicalization tells search systems which page is preferred among similar versions. We verified these mechanical rules, not Google's final search decisions.

For implementation guidance, read the hreflang SEO guide or inspect a site's discovered XML sitemaps with the free sitemap checker. The sitemap checker is a separate tool; it does not reproduce this study's 16-language reciprocal-link audit.

For perspective, the 2025 HTTP Archive Web Almanac reports hreflang markup on about one in five pages in its much broader web dataset. That is a different population and measurement method; it must not be compared directly to our deliberately selected multilingual pages.