AI Search & GEO Glossary

GPTBot

Definition

GPTBot is OpenAI's web crawler, identified by the GPTBot user-agent, used to collect publicly available web content that may improve OpenAI's models. Sites control it via robots.txt: allowing GPTBot makes content available for model training and related uses; OpenAI also operates separate agents (like OAI-SearchBot for ChatGPT search) with distinct user-agent strings and purposes, per OpenAI's published bot documentation. For GEO, the practical decision is which OpenAI bots to allow: blocking GPTBot while allowing OAI-SearchBot keeps you out of training data but visible in ChatGPT search results. Verify real GPTBot traffic before drawing conclusions from server logs: OpenAI publishes its IP ranges, and spoofed user-agents are common. The robots.txt decision is also reversible — many publishers allow search-facing bots while restricting training crawlers, and revisit the split quarterly as OpenAI's bot roster changes.