Pinterest offers Lens, a camera feature inside the Pinterest app that lets any user point their phone at a real-world object and find visually similar Pins and shoppable products. The feature launched in beta in February 2017 and was upgraded in May 2025 to support multimodal search, powered by visual language models, initially for women's fashion in the US, Canada, and the UK. The system can identify approximately 2.5 billion objects across home and fashion and handles hundreds of millions of visual searches each month.

Pinterest Lens uses deep convolutional neural networks trained on Pinterest's image corpus to identify objects, their attributes, and their visual style. The May 2025 update added visual language models (VLMs), a form of generative AI, enabling the system to process combined text-and-image queries; for example, a user can tap a Pin and add a text keyword such as a color or style to narrow results. An animated glow feature, initially available for women's fashion in the US, Canada, and the UK, uses VLMs to generate descriptive words that help users identify what they like about a particular Pin. A separate Shop the Look feature uses the same computer vision infrastructure to identify and link individual shoppable items within a single lifestyle photograph.