LongCat-Video (demo: https://huggingface.co/spaces/multimodalart/LongCat-Video and documentation: https://github.com/meituan-longcat/LongCat-Video) generates a 6-second video from a text instruction, or an image with an instruction.

Lightweight model and specially trained for generation from an image source.

On Hugging face, choose the resolution, the “seed” and the generation mode (fast or high quality).

The full version of the template, once installed locally, can generate long videos without the limitations of the demo (requires technical skills and a proper computer).

Free and open-source.

Image for test: https://commons.wikimedia.org/wiki/File:Finnish_Air_Force_Hawker_Hurricane_warbird_and_US_Navy_T-6_Texan_warbird_in_flight_over_Finland.jpg