ChatGPT Search
A ChatGPT feature from OpenAI that combines conversational AI with real-time web search.
ChatGPT Search is a feature of OpenAI's ChatGPT platform that combines the capabilities of a large language model with real-time web search. Unlike the original ChatGPT, which answered only from its training data, ChatGPT Search actively looks up current information on the internet and integrates it into a conversational answer.
How ChatGPT Search works:
- The user asks a question (for example "best electronics online stores in Slovakia")
- ChatGPT analyses the question and decides whether it needs current web data
- If it does, it runs a web search through partnerships with data providers
- It integrates the search results into a conversational answer with source citations
- The user gets an answer with direct links to the cited pages
What matters for online stores:
ChatGPT Search indexes the web with its own crawlers, identified by the user-agent "OAI-SearchBot" (for real-time search) and "GPTBot" (for training data). For online stores it is critical that these crawlers are not blocked in robots.txt.
Differences from Google Search:
- ChatGPT Search does not rank results as a list of links; it combines information from several sources into a single answer
- The user can ask follow-up questions within the conversation
- ChatGPT Search can compare products, recommend alternatives and explain differences
- Answers are personalised based on the context of the conversation
Model used for search: ChatGPT runs on the GPT-5 family; in search mode it combines the model with live web search (Pozorio measures via gpt-5.6-luna + the web search tool). In 2026, ChatGPT has more than 200 million weekly active users.