Mochi

Mochi 1 is an open-source text-to-video generative AI model developed by Genmo. It uses a diffusion-based architecture designed to generate short, high-quality videos from natural-language descriptions while maintaining reasonable motion and visual consistency. Unlike a typical online video-generation application, Mochi is primarily a model and developer framework that can be downloaded and run locally or integrated into AI interfaces such as ComfyUI. It is particularly useful for developers, researchers and advanced AI users who want control over the video-generation pipeline rather than relying entirely on a closed cloud service. The model is released under the Apache 2.0 license, allowing commercial use subject to the license terms. Country availability: No specific country restriction is publicly stated for the open-source model. Local use is generally possible internationally, although access to model-hosting platforms, cloud GPU services or third-party integrations may vary by country.