@RRulehard@shortformsavage you can do it from scratch. Source: I did this multiple times and pages have 100m plus views. But algorithm has greatly changed in the last few months
This IG profile is 1 month old and getting millions of views using this workflow 👇
1/ download the video (snapinsta)
2/ extract a clean frame (any frame tool or locally)
3/ drop the frame into chatgpt → "deconstruct this into a detailed prompt. focus hard on framing, background, setting, characters, and pose"
4/ new chat → remake the photo but put YOUR clothing on them
5/ take that image into higgsfield / replicate / wavespeed (wherever you access Seedance 2.5)
6/ write a seedance 2.5 prompt based on the reference video. (guide i follow: https://t.co/XCAitdHHJt)
@ClaraV86195 @ecomrudolfs@maxxmalist you can upload your own voice Here: https://t.co/aOdC4CSKap
as far as being limited to 10 seconds, yes it is annoying. You have to do multiple cuts. But for lipsync it is definitely the best model I've used. Results are much better than infinite talk IMO