First data: llms.txt is being fetched, but not specifically by AI bots

A look at the Achtung.app server logs reveals a clear pattern: llms.txt does get fetched, but not by the AI bots the file is meant for.

In the 30 days leading up to 17 May 2026, Achtung.app recorded 90 requests for "/llms.txt". The breakdown:

• 76 requests from a single scraper using a spoofed Chrome user-agent (one IP)
• 2 requests from the SEO crawler Barkrowler (babbar.tech)
• 2 requests from our own AchtungBot (internal probes)
• 2 requests from a spoofed Chrome 86 user-agent, 2 from a spoofed Pixel 6 mobile user-agent
• 1 request each from AhrefsBot, BuiltWith, Dataprovider and three other profiling user-agents

ClaudeBot, GPTBot, PerplexityBot, OAI-SearchBot, Claude-SearchBot, ChatGPT-User, Perplexity-User, Applebot, Googlebot and Bingbot did not appear in the list once, even though those same bots otherwise hit Achtung.app over 17,000 verified times in the same window.

This lines up with Google's official AI optimization guide: per Google, special files like llms.txt have no effect on AI visibility. The clients fetching the path are classic SEO and discovery crawlers that pick up every documented ".txt" file.

Achtung.app continues to serve an llms.txt as a declarative entry point for agents that explicitly request that path. The file costs nothing and does no harm. It is just not a lever for more AI citations.

Update, 17 May 2026: A few hours after this note was published, GPTBot fetched the file for the first time. The request was one URL in a normal sitemap-driven crawl of ten pages within eight seconds (alongside "/methodik", "/datenschutz" and "/marken"), not a targeted llms.txt fetch. That reinforces the picture rather than overturning it: llms.txt gets picked up because it sits in the sitemap, not because it is treated as a special AI signal.