DynamiCrafter is a text-to-video model capable of generating approximately 2-second dynamic videos based on input images and text. This model, trained to generate 576x1024 resolution videos, excels at capturing the dynamic effects of both input images and text descriptions, producing realistic short video content. Suitable for video production, animation creation, and other video generation scenarios, DynamiCrafter serves as a powerful productivity tool for content creators. The model is currently in the research stage and is available for personal and research use only.