# Robots.txt for Vernaestivus Garden & Interior Design Barcelona
# https://vernaestivus.eu/robots.txt

# Allow all search engines to crawl the entire site
User-agent: *
Allow: /

# Explicitly allow important pages and sections
Allow: /#services
Allow: /#portfolio
Allow: /#contact
Allow: /favicon.svg
Allow: /favicon-32x32.png
Allow: /favicon-16x16.png
Allow: /apple-touch-icon.png
Allow: /android-chrome-192x192.png
Allow: /android-chrome-512x512.png
Allow: /site.webmanifest

# Allow language-specific URLs
Allow: /?lang=es
Allow: /?lang=de
Allow: /?lang=ru
Allow: /?lang=en

# Block unnecessary files and directories that could waste crawl budget
Disallow: /node_modules/
Disallow: /src/
Disallow: /.git/
Disallow: /.vscode/
Disallow: /.bolt/
Disallow: /dist/
Disallow: /build/
Disallow: /*.log
Disallow: /*.json$
Disallow: /package-lock.json
Disallow: /tsconfig*.json
Disallow: /vite.config.ts
Disallow: /tailwind.config.js
Disallow: /postcss.config.js
Disallow: /eslint.config.js

# Block duplicate Google verification files
Disallow: /google7dadfa775a55440e*.html

# Block temporary and backup files
Disallow: /*~
Disallow: /*.tmp
Disallow: /*.bak
Disallow: /*_backup
Disallow: /*.old

# Block common admin and development paths
Disallow: /admin/
Disallow: /wp-admin/
Disallow: /wp-content/
Disallow: /wp-includes/
Disallow: /cgi-bin/
Disallow: /tmp/
Disallow: /temp/

# Block search and filter parameters that could create duplicate content
Disallow: /*?*utm_*
Disallow: /*?*ref=*
Disallow: /*?*source=*
Disallow: /*?*campaign=*
Disallow: /*?*medium=*

# Allow specific search engine bots with enhanced permissions
User-agent: Googlebot
Allow: /
Crawl-delay: 1

User-agent: Bingbot
Allow: /
Crawl-delay: 1

User-agent: Slurp
Allow: /
Crawl-delay: 2

User-agent: DuckDuckBot
Allow: /
Crawl-delay: 1

User-agent: Baiduspider
Allow: /
Crawl-delay: 2

User-agent: YandexBot
Allow: /
Crawl-delay: 1

# Block problematic or aggressive crawlers
User-agent: AhrefsBot
Disallow: /

User-agent: MJ12bot
Disallow: /

User-agent: DotBot
Disallow: /

User-agent: SemrushBot
Crawl-delay: 10

User-agent: MajesticSEO
Crawl-delay: 10

# Sitemap locations
Sitemap: https://vernaestivus.eu/sitemap.xml

# Host directive (helps with canonicalization)
Host: https://vernaestivus.eu