Discover Your Style with Doppl (techlife.blog)

Google Labs on June 26, 2025, launched Doppl, an experimental AI/AR app that creates dynamic virtual try-ons by synthesizing videos of a user “wearing” outfits sourced from social media, friends, or retailers. Building on Google Shopping’s recent virtual-try-on scale-up, Doppl generates a digital representation of a user and maps uploaded clothing images onto that avatar, letting people preview, save, and share looks on iOS and Android in the U.S. The emphasis is on exploration—trying new styles without physical fitting rooms or returns—and on social feedback to help users refine personal style. For AI/ML practitioners, Doppl is notable because it combines body-conditioned image/video synthesis, garment transfer, and AR rendering at consumer scale. Technically this implies model work on robust pose and shape estimation, photorealistic texture transfer, temporal consistency in generated video, and integration with large item catalogs. The experiment also highlights real-world implications: potential to boost e-commerce conversions and personalization, but also challenges around imperfect results, fairness and fit estimation, dataset bias, and privacy/deepfake concerns. Researchers and engineers will need to address fidelity, safety, and consent mechanisms as virtual try-on systems move from lab demos to mainstream shopping tools.