i think i found a way to outsmart pika labs. I just grab the first second of the vid, then i recreate another inflation video with that frame, and so go on until you're satisfied. I also made this to showcase it

Back with a few more. The bikini ones came out way better than expected.

More

More

More

>>27366 Based and architecture pilled

>>28420 Hey, OP of these images here... nice

>>28468 Oh nice. Love your work. Including your cgi vids.

what are the prompts to make these expansion vids?

>>28151 You've got some great stuff on your dev page. Hope you don't mind i had to make some wg animations from your Daenerys pic

I'm very happy with how this one came out

>>28532 If you're referring to these: >>28415 >>28416 >>28418 >>28420 For kling I tend to use some version of "X suddenly gains a massive amount of weight. Her buttocks expand out behind her and grow fatter. Her hips and thighs expand and grow fatter and wider. Her breasts expand out from her bosom and grow fatter. Her movements should be smooth and natural." And change it based on the pics. For luma I use "chest" instead of breasts to get around the censor.

still works great with before + after pics

If you use before and after images with Vidu, it generates peak videos, i posted one on the fake inflation scenes thread on the inflation catalogue

Can someone pass me the links of all the Ai websites you guys use? Kling doesn't work for me.

>>28720 First: you need also a end image keyframe or else the video will not work Second: Vidu honestly works very well with keyframe images I use Leonardo AI, Flux Org or LTX Studio for get my images, but i would like to know if there is any other better ai to do that

My contribution to this theme. More like a proof of concept at this point. The images i used to make video were really shitty and rushed. Was trying to make a consistent weight gain story using only AI technology and a little bit of editing.

Another one

>>28418 Damn, AI is based, actually

Runway is a bitch in terms of censor, limited generations and blackouts but you can't argue with results.

>>28805 Holy shit, is that with before and after images? I haven't touched Runway since it doesn't replenish free credits....

>>28805 Very nice! I would LOVE the eating ones if the AI could figure eating out, but because it can't I actually prefer the one that's just expansion without the eating.

>>28809 Yes. The after image is from pina18colada on dev and i made the before image using stufferai, and open ai for a similar pose and look. Face swapped the face to make them look like the same person. Changed the color of the befores dress, then put the before in the afters background. Attached pics sorta show my dumb process. And yeah it's insane that you get 2 tries with runway then you're done for the month. Even with my 30 dummy accounts I ran through that quick lol >>28810 Same. I've tried adding a funnel, specifying different food items. Never works well lol

The results from kling were good too. First sequence where I didn't get a single bad result lol.

Another recent attempt. Wendys calls it the Krabby Patty for a reason

someone tried to use Vidu's keyframe generator? It can make great stuff. I also used LTX Studio to make the images, cuz technically the ai was made to make images for generate videos...but the way it generates videos is lame

Can you please share your prompts friend :O

>>28818 "Cinematic, dramatic, final fight, a beautiful woman, her body inflates imto a huge and obese rubber woman, angry expression, full transformation on camera" But honestly what makes the magic isnt the prompt alone, its the images, you need an image for the start and the end. Cuz Vidu is great to render inflation motion...but not so great to make designs... So by using images Vidu skips that task and makes This

>>28815 Omg

Anyone have any tips to get past Hailuo censors? They've gotten real strict all of a sudden

So for 2 image video generation I have: Luma Kling Vidu Runway Hapier Does anyone know of any others? Kling and Luma have gotten pretty stingy lately with the load times.

>>28876 Also a very good question would be. How can u make 2 ai images sequences easily, that looks like the same person with the same clothing? Btw i can't use ai retouch cuz is for phone ;-;

here's a video i made using Vidu, LTX Studio and some editing , for showcase what we can do (i send the same vid on the fake scenes made by ai thread i made on inflation catalogue)

I tried generating something out of some fantastic generations I had on civit, and I think they went well. 4 seconds isn't enough tbh and the faces could be better but holy damn

Someone knows any good inpainting ai?

>>28933 you did well with these

aparrently PixVerse makes great expansion Art, atleast thats what i found on deviantart

>>28996 The moderation is heavier than Hailuo though.

>>29094 What? What are your prompts? Maybe youre being too obvious?

My prompts normally are stuff like "Fantasy Action movie, on X location, a woman wearing x thing, she swells up and expands until she is a huge and obese inflatable woman" or "Her body inflates, her belly inflates, her arms inflates, her face cheeks inflate"

Looks like Pixverse can do weight gain out of the box, at least with a starting image. Just told it "The woman becomes fatter."

Looks like Pixverse seems to be the easiest one. Made two with little more than a picture and some basic prompts seems to be really good.

Miss Bella Brookz joins the BBW Community

Took a twitter girl's weight loss journey pics and reversed them using vidu.

Oh my god, Vidu's "Reference to Video" feature is amazing. The first two weren't made with it, but the other two were (don't ask about the sudden jump in quality, vidu does that when you make an account now). I wonder what else this Reference thing can do?

Do you think you could share some of those please friendo? I keep getting weird results and I was hoping for some advice.

Some more from Pixverse.

>>29531 Prompt?