# Robots.txt for The Digital Solutions
# Last updated: August 3, 2025

# Allow all web crawlers to access all content
User-agent: *
Allow: /

# Disallow common admin and system directories (if they exist)
Disallow: /admin/
Disallow: /wp-admin/
Disallow: /wp-content/uploads/
Disallow: /cgi-bin/
Disallow: /tmp/
Disallow: /logs/
Disallow: /_private/
Disallow: /error/
Disallow: /stats/

# Disallow specific file types that shouldn't be indexed
Disallow: /*.pdf$
Disallow: /*.doc$
Disallow: /*.docx$
Disallow: /*.xls$
Disallow: /*.xlsx$
Disallow: /*.zip$
Disallow: /*.rar$
Disallow: /*.exe$

# Allow important files for SEO
Allow: /sitemap.xml
Allow: /robots.txt
Allow: /favicon.ico
Allow: /*.css
Allow: /*.js
Allow: /*.png
Allow: /*.jpg
Allow: /*.jpeg
Allow: /*.gif
Allow: /*.svg
Allow: /*.webp

# Crawl delay for polite crawling (optional)
# Crawl-delay: 1

# Specific instructions for major search engines
User-agent: Googlebot
Allow: /

User-agent: Bingbot
Allow: /

User-agent: Slurp
Allow: /

User-agent: DuckDuckBot
Allow: /

User-agent: Baiduspider
Allow: /

User-agent: YandexBot
Allow: /

User-agent: facebookexternalhit
Allow: /

User-agent: Twitterbot
Allow: /

User-agent: LinkedInBot
Allow: /

# Block bad bots and scrapers
User-agent: AhrefsBot
Disallow: /

User-agent: MJ12bot
Disallow: /

User-agent: DotBot
Disallow: /

User-agent: SemrushBot
Disallow: /

User-agent: BLEXBot
Disallow: /

# Sitemap location
Sitemap: https://thedigitalsolutions.net/sitemap.xml

# Additional sitemaps (if you create them)
# Sitemap: https://thedigitalsolutions.net/sitemap-images.xml
# Sitemap: https://thedigitalsolutions.net/sitemap-news.xml