This workflow allows you to create long videos and also supply a reference face to hopefully improve character consistency.

The character face is replaced on each 5 second video segment, to try to maintain facial features. May not works if character is not looking at camera.

Prompts can be supplied for each video segment on the long video.

You will need to make sure comfy has been updated to support "Subgraphs"

This workflow uses subgraphs

WAN 2.2 LIGHTNING

https://huggingface.co/Kijai/WanVideo_comfy/tree/main/Wan22-Lightning

GGUF

https://huggingface.co/QuantStack/Wan2.2-I2V-A14B-GGUF/tree/main