Pinterest Enhances Visual Search with New AI Tools and User Refinements

Pinterest is expanding its visual search functionality with a series of new features aimed at making image-based discovery more precise and accessible. Announced on Monday, the update introduces AI-driven tools designed to help users better understand and refine what they’re looking for – especially when that “something” is hard to put into words.

The most notable change involves the integration of visual language models (VLMs) to help explain why a user is drawn to a specific Pin. Initially rolling out in the women’s fashion category across the U.S., U.K., and Canada, this feature uses generative AI to surface descriptive keywords when a Pin is tapped, offering deeper insight into personal preferences. The visual feedback appears as a subtle animated glow, and the AI-generated descriptors then help narrow future searches.

Another new capability allows users to fine-tune fashion-related results by seeking variations of an item, such as different colors, styles, or fabrics. For example, a user might search for a similar jacket that better fits a formal setting or explore an outfit with a more Y2K-inspired look.

Pinterest is also expanding where visual search can be initiated. Users will now be able to long-press on any Pin in their Home Feed to begin a visual search, improving accessibility across the app.

These updates arrive shortly after Pinterest’s decision to begin labeling AI-generated images and offer controls over their visibility, following user feedback that AI content was making authentic image discovery more difficult.