Genmo
Free tierOpen-source video world models that understand the physical world like never before
Free tier available·All audiences·Powered by Genmo (Mochi 1)·Open source
Key strengths
State-of-the-art open-source text-to-video model (Mochi 1)Fully self-hostable and locally runnableComfyUI integration for no-code customizationInteractive playground for prompt-based video generationResearch-grade model with physical world understanding
Free tier + paid plans
US
Self-hostable
No ratings yet
Genmo develops state-of-the-art diffusion-based video generation models, with Mochi 1 being their headline open-source release. The model is distributed via GitHub and HuggingFace, and can be run locally using a provided quickstart shell script and Python CLI. It supports integration with ComfyUI for node-based workflow customization. The architecture is built around understanding physical world dynamics for realistic motion and scene generation, positioning it as a research-grade foundation model for video synthesis tasks.