Day 75: What’s the HARDEST part of making AI videos?

Day 74: workflow! this is the reason you spend days/weeks on one ai video since i got a workflow ai video generation became way easier for me. steal my workflow: 1/ idea/story write out the story/idea you want to create, you can write this inside of chatgpt or claude and ask it to find loophole or give suggestions. (you can always make edits to your story) once the story is up to your taste, move to the next part. 2/ scripting this covers more details about your story like: > scenes > camera angles > characters + environment > the action of each character this is where you dictate how the video would play, be creative about it. you can also ask gpt or claude to write it, then proof read and make corrections to ensure it's in line with your idea/story 3/ character generation after identifying my characters and environment, i move to midjourney to create them. i use mj because it gets my style correctly but it makes silly errors sometimes so i use gpt 2 to correct those errors. (dm for my correction prompt) 4/ video generation for this part i use "seedance clean" it's a claude skill that writes properly detailed seedance skills. it reads your script + characters and writes prompts covering: > audio > optics > camera > lighting > art style > wardrobe > performance > positive locks > character action repost and comment "skill" i'd send it to you proof the prompt for errors and feed it to seedance for the video generation. most times the videos come back with errors, don't panic, lol just rewrite the prompt to fix those errors and regenerate 5/ edits here to join the videos together, add background sounds/music, captions and other effects to make the video look better. personally i use davinci resolve but there are many other tools like capcut and adobe premiere. that's all... hope this helps!❤️

Day 73: short commercial for @0xfluid