Officevibe — AI visibility report

Methodology. Data extracted from nginx access logs filtered by AI-agent user-agent strings against a maintained crawler manifest list (OpenAI ChatGPT-User & GPTBot & OAI-SearchBot, Anthropic ClaudeBot, Meta-ExternalAgent, Perplexity Bot & Perplexity-User, Google's GoogleOther & Google-Extended, Apple's Applebot-Extended, You.com YouBot, ByteDance Bytespider, Common Crawl CCBot, Mistral, DuckDuckGo DuckAssistBot, others). The pipeline runs daily at 01:05 UTC and writes structured JSONL to /var/log/nginx/llm_crawlers.jsonl. This report aggregates the prior Mon–Sun window. Live retrieval bots fetch in response to a real-time user query (e.g., a person asking ChatGPT or Perplexity a question that triggers a search). Training/index bots crawl on the operator's schedule for model training or search indexing. What this does not measure: retrieval-augmented generation events that don't fetch from this URL, fetches that bypass the crawler-manifest UA strings, or what AI agents do with the content after retrieval. Report regenerated every Monday at 02:00 UTC; numbers shift with each new window.