text2vid img2vid H3 Minimax - Fully uncensored local AI video generator released today

I know that there's a reference to vid model and we can use video with it, so I wonder if we can do a nudify workflow from this? like make the video of a dancing girl get nudified?
 
it works just by default without any special workflow just by prompting: "naked female dancing/garment falls off revealing her...". Better results are achieved if you give it a reference image of a naked body so it basically does a swap of what you want.
 
in comfyUi I can run only 5 times is their any way to get multiple run free?
 
there's no separate audio prompt, you can describe the audio in the normal prompt and/or connect a "Load audio" node if you want a reference. (same voice/effects and so on)
 
did you find anyway to bypass 5runs from comfy? Any email trick or any other guide?
 
just type a random email, they don't check if it exists.
works fine on my side.
 
but are you able to do that with a free account? if I click "Import" it prompts me to subscribe
 
you can, for the moment. The check for that is client side so it works even on free accounts.
 
I can't believe that...they honestly have ZERO checks server side, I literally typed a RANDOM email that doesn't event exist and I was able to register...damn vibecoders
 
Are you able to nudify a person using the exact reference video? Like just taking the clothes pff but it is the exact same video. Like a tiktok video? Or is there any workflow other than h3 that is capable in doing this?
 
Going the video route will still be better because the character will stay consistent throughout the frames.
 
Back
Top