Genmo

Genmo is a research lab focused on advancing open-source video generation models, particularly through its flagship model Mochi 1. The core function of Genmo is to transform textual descriptions into high-quality video content using a text-to-video generative AI system. Users provide a written prompt—such as "a slow-motion shot of a glass shattering on the floor, pieces scattering" or "a time-lapse of a city street artist creating a chalk mural on the sidewalk"—and the model produces a corresponding video that captures motion, spatial detail, and temporal dynamics. The platform is designed to serve as a playground for both casual experimentation and serious creative work, allowing users to generate videos directly from the website’s interface by entering prompts and clicking a generate button. The underlying technology, Mochi 1, is an open-source model that can be downloaded, run locally, and customized by developers and researchers. This openness enables fine-tuning for specific use cases, integration into larger pipelines, or academic study of video generation architectures. Genmo’s research mission extends beyond simple video creation: it aims to build "world models" that simulate the physical world, including object interactions, lighting changes, camera movements, and realistic physics. By modeling how objects break, how water flows, how characters move, or how environments evolve over time, Genmo seeks to contribute to the broader goal of artificial general intelligence (AGI) that understands spatial and causal relationships. The platform’s featured prompts demonstrate this capability: a theater stage shot from behind curtains showing actors preparing, or a close‑up of a dragon’s eye opening with iridescent scales. Each example tests the model’s ability to handle