GAEO.ru moved from Tilda to Astro and GitHub Pages with bilingual pages, reusable templates, industry content and release QA.
My GAEO.ru project needed more than a visual redesign. It had to retain existing URLs while growing into a structured library of industry landing pages, technical methods and verifiable case studies.
The original website ran on Tilda. As more methods and case studies were published, shared components, language versions and content links became harder to maintain manually.
The new information structure connects industry landing pages, methodologies, editorial content and Alexey Yakovlev's biography. Each content type has a distinct evidentiary role.
Astro generates the HTML hosted by GitHub Pages. Canonical source files and components live on astro-source, while main stores verified published output. Shared navigation, footer and calls to action come from central components. Existing HTML pages coexist with Astro templates in a managed hybrid architecture.
Core content is available without client-side JavaScript. Yandex Cloud Functions and Postbox handle the separate enquiry flow.
Every new indexable page is checked for its title, meta description, single H1, canonical URL, mutual hreflang for real translations, Schema.org and internal links. Commercial pages explain specific industry needs; methods and cases offer separate supporting evidence.
Published examples include e-commerce and industrial B2B. Structured data alone cannot guarantee AI recommendations.
Changes go through pull requests, Astro builds, source preservation and browser tests at 1440, 390 and 320 pixels. The tests cover images, the lead form and horizontal overflow. Production is deployed to main only after the source checks succeed.
In the October 26, 2026 release, 70 browser checks passed. QA reported 67 HTML files and 62 sitemap URLs in that release. These numbers do not demonstrate search-engine indexing.
I define business positioning, information architecture, editorial rules and acceptance criteria, using coding tools and automation for implementation. I do not claim to have manually written every source file. The confirmed result is a deployed site with documented release checks.
The separate AI visibility case reports an increase from 22% to 48% in an earlier period. That growth cannot be attributed to the later platform migration.
Start with the site's business purpose, content model and technical constraints.
