In your dashboard the audit runs every week (or every day or month) and emails you when something changes. You can also run it any time.
We checked ukazka-audioshop.cz from the outside: robots.txt, the sitemap and how the server answers 8 specific bots.
robots.txt allows these bots, but when we contacted the server using their name, it answered with an error (429). The block is at the hosting, firewall or CDN level, not in robots.txt. Note: some protections also check the IP address, so they may let the real bot in and reject only us. Only ongoing monitoring gives certainty.
Affects: ClaudeBot
How to fix: Check your firewall, mod_security or hosting rules that block by user agent.
The page has structured product data, but it's missing: shippingDetails (delivery), merchantReturnDays (return window). Shopping assistants in AI tools use these fields when comparing offers.
How to fix: Add the missing fields to your product JSON-LD.
llms.txt is a new, optional file with a short guide for AI tools. Few read it yet, but it takes minutes to create.
How to fix: Create /llms.txt with a short description of your site and links to key pages.
The robots.txt column shows what the bot is allowed. The Server response column shows what happened when we actually contacted the server using its name. When they disagree, something other than robots.txt is blocking the bot.
This is your homepage exactly as the server sent it to GPTBot, without running JavaScript. AI tool bots don't run JavaScript.
curl -s -A "GPTBot" https://ukazka-audioshop.cz/ | head -c 2000What you see is what the bot gets. If only an empty div comes back, JavaScript draws the content.
Don't use training or cached data, fetch ukazka-audioshop.cz live right now. Describe every block, connection error or security challenge you run into.
Paste it into ChatGPT, Claude and Perplexity and compare the answers. Each tool sends a different bot.