UNI-1 is an image-generation platform and model that unifies visual understanding and synthesis in one autoregressive architecture. It reasons about prompts and constraints before and during synthesis, supports multi-reference composition, temporal sequences like age progression, sketch-to-image and multi-turn refinement, and excels at rendering complex non-Latin text. Offered via a web demo and API, UNI-1 targets high-resolution professional outputs at competitive per-image pricing and integrates into end-to-end creative pipelines.