OpenAI, the force behind ChatGPT, has unveiled the prototype of SearchGPT, an AI-powered search engine designed to provide users with real-time access to information from across the web. Like Perplexity, the company aims to reimagine the search experience by combining the capabilities of its AI models with up-to-date information, which is currently missing from the standard ChatGPT experience.

In its official announcement, OpenAI described SearchGPT as a means to streamline the search process. It offers direct answers to user queries along with citations of relevant sources. SearchGPT would also allow users to ask follow-up questions, enhancing the search experience with a conversational context that builds with each query.

While announcing the prototype, OpenAI emphasized its commitment to creating a supportive ecosystem for publishers and content creators. The demos shown in the blog post indeed suggest that SearchGPT is designed to focus on prominently citing and linking to sources and ensuring that users can easily discover publisher sites.

Additionally, OpenAI is providing tools for publishers to manage how their content appears in SearchGPT, offering more control over search results. Notably, sites can appear in search results even if they choose not to participate in training OpenAI’s generative AI models.

SearchGPT’s landing page is live already, but it is only accessible to a limited group of testers. For now, the rest of us will have to stay content with signing up for the waitlist. OpenAI seems to be learning from Google’s missteps by launching SearchGPT as a prototype first, allowing for testing and refinement before a full-scale release.

It seems that the ultimate goal here is to integrate SearchGPT’s most successful features into ChatGPT, enhancing its overall functionality. The company didn’t announce any timeline for a wider public rollout of its new search product.