ChatGPT launches shopping research tool for smarter product discovery

OpenAI has introduced a new shopping research tool in ChatGPT that helps people find the right products without browsing multiple websites. Users can describe what they need, such as a quiet vacuum, a laptop under a certain budget, or a gift idea, and ChatGPT creates a personalized buyer’s guide in minutes. The feature asks helpful clarifying questions, searches trusted online sources, reviews product details, and uses past conversations and ChatGPT memory to refine recommendations. It is now rolling out on mobile and web for ChatGPT Free, Go, Plus, and Pro users, with nearly unlimited usage available through the holiday season.

Shopping research is designed for deeper decisions where people want to compare features, check trade-offs, and understand their options. It works especially well for detailed categories like electronics, beauty, home goods, appliances, and outdoor gear. For simple questions like checking a price or confirming a feature, a standard ChatGPT reply still appears quickly.

Users can search for new products, find similar items, compare choices side by side, browse deals, or get gift suggestions. ChatGPT Pulse, available to Pro users, can also recommend personalized buying guides based on past chats, such as suggesting accessories if someone has been discussing e-bikes.

To use the feature, users can start with a shopping-related question or select “shopping research” from the tools menu. ChatGPT then opens a visual interface where users can explain their needs, answer questions about budget and preferences, and guide the research by marking products as interesting or not relevant. The tool searches the internet for up-to-date prices, specs, availability, reviews, and images, then presents a summary of the best options and their key differences. Purchases can be made by visiting the retailer’s website, with direct checkout expected in the future for supported merchants.

The system is powered by a version of GPT-5 mini trained with reinforcement learning to handle complex shopping tasks. It reads reliable retail sites, cites sources, and adjusts its recommendations in real time based on user feedback. OpenAI says chats are not shared with retailers, and results come from reading public product pages, not paid placements. The company notes that pricing and availability may occasionally be inaccurate, but it plans to keep improving accuracy and expand support across more categories.