These AI tools add Working 10 step Hunyuan Video portrait to Hunyuan. 640x480.

3 - 7 credits. Usually 5.5. 2 seconds.

It's uncensored. I have quite a good deal of uncensored videos if you want to turn on NSFW and look at my posts so you can get a handle on the prompting. So far I've gotten BJs and FJs and cowgirl and doggy and prone bone and walk in and armpit jobs and ass jobs and tit jobs and body surfing nude/sex and wetlook and the pose. Threesomes and orgies work too. Taking off shirts, unbuttoning and opening shirts works but you might just get flashed, taking off bottoms, shoes, socks, does not work for anybody. Nor does pulling blankets off.

Nuns work fine, uniforms with epaulettes, Star Trek uniforms, superhero uniforms, clown makeup, any color skin works, but if you prompt demons they look a bit more cartoonish than if you didn't. Pointy ears just works though. Speaking of cartoonish, any time I prompted anything cartoon or anime I was disappointed, save once. It's very vintage anime, though I'd go anime again before I went cartoon, cartoon is crap. Riding bikes works, women on car hoods, running, car/train/airplane interiors, in the rain. Sniper rifles, hand guns, blood. Sniper view (CUT TO: The view now appears to be shot through a sniper scope of:)

It knows far, far more locations than Pony. Anywhere you want, really. Golden hour looks nice. Sun peeking through the whatever looks nice.

I can't get it to do cunnilingus or boob sucking or boobs against glass or legs behind head or shoe dangle or full nelson position or piledrive position (though it does upside down fine) or suspended congress or female masturbation. Or a woman diving which I have tried like ten times. Please let me know if you prove me wrong. I thought it couldn't do the man's face and the woman's face during a sex act, but it can, it's just nearly impossible.

The feet need a "She has 5 toes and 5 toes alone." It likes to show soles when you ask for feet so plan for that. It won't show stockings with no shoes unless you add socks. But black stockings I haven't gotten even with socks, it's high heels the whole way down.

I hear people saying that they get still images. I avoid still images by giving it an entire sentence of direction that is just about the motion. I still do get them say 5-10% of the time.

I use FLUX style "hanging" where you use full sentences and describe the entire scene using stage directions, like "The woman on the right hand side is dancing with her hands in the air wearing a halter top and short jean shorts, in the middle along the top an airplane in flying across the sky in the background is pulling a sign that says "HUNYUAN VIDEO", on the right is a seated man wearing clown makeup and a sweater and jeans and loafers with his head in his hands is looking sad."

I hear that you should try to keep it down however because it just stops reading if you make it too long. I try to keep it less than 300 words, IDK.

Text, I have seen work only once. Usually, it is unreadable.