Guide Fake img2vid How to make NSFW videos with your photos for free using Tensor.art? Guide to Wan Image to Video (i2v) tool with LoRA. Promts included.

[megaton123]

Ok, many questions about faceswap and BJ videos. Mostly i'm ignoring single questions as most of them there is already answer in thread but when there is almost a page with the same requests, here we go again :D

What is the best way to do a face swap on a picture? I have two pictures and want to swap the faces. I read that the faceswap lora's are not good. I tried this one https://tensor.art/template/906922065750186706. It turned out terrible. I thought I could do it in the img2img in workflow but i can't get it to work. Any suggestions?

Can you share the loras and prompt you use for BJ? I tried to use https://tensor.art/models/901903731472234580 which i believe is the same lora used in the DEEP BLOW2 template, but I cant seem to get a proper looking dick, eye contact, and head movement like in the template.

Also, is there a way to 100% replicate a template on workspace using the same lora, like extracting whats in the template and replicate it? I've only been looking into ai stuff couple of days ago, and has only been experimenting with loras with wan 2.2, so I'm not familiar at all with all these. I'd really appreciate if you can help with this.

Spoiler

For missionary I've been able to get some good results using workspace and the Sex FOV slider lora

So i've used 2 images of 2 Ai babes. One of her has small tits which you don't like but lovely blue eyes. Second has huuuge tits but you don't like her face and brown eyes. So you're taking body you like and face you prefer and just delete brown eyes girl face leaving white background. Then you paste face of blue eyes girl. You can make it probably in many tools, free GIMP for example.





Then you have something like that:



There is answer for questions about workspace, prompt for BJ and lora.


Only 6 steps with 1 cfg, 6 seconds 23 credits = so it's not even expensive, probably even cheaper than ai tool and result is still better.

Finally, video:

And final addnotation: As you can see in prompt there is "Woman lands on her knees" - she wasnt fall on her knees and camera moved. That's because she didn't have to, if source image would be her full body with legs, she would fall on her knees and camera would zoom in into action. Camera in oral insertion lora always has to be close to action because it was trained to do it.
Also not every lora can make face swap, some of them sucks.
Also "Make sure to mix body and face" it's not needed AI would do this anyway. But sometimes it could be helpful.
Lora's for BJ:

https://tensor.art/models/901034713559385392

https://tensor.art/models/900201419036968596

 

[tiktokzengin2]

how to make this, propt and loras ?

 

[Kickassandedge]

This might be a sign for me to stop Gooning and get my shot together

 

[erinnatsuki]

Still making use of Tensor Art for the templates, a lot of them still work. However, I decided to focus more on making full-length scenes instead of clips, so I also made use of Seaart.Ai and Akool instead for most of this video, especially with establishing scenes, transitions and face swaps. Haven't migrated to TH yet, especially since I feel that TA is still usable for many of the templates floating around and, yes, I still utilize the free trials with alt accounts (simply because I'm already paying a lot for Seeart and Akool LOL). Below are the templates I used on TA:

Pose Change: https://tensor.art/template/928931838259363850
For Doggy and Missionary (standing, haven't tried with on-the-bed photos): https://tensor.art/en-US/template/938312367882694493
Blowjob 1: https://tensor.art/en-US/template/916556090785681009
Blowjob 2: https://tensor.art/template/904080795633741558?X-Device-Id=Ba6rwqxzbTSU9kQK0Jaa6
Facial: https://tensor.art/en-US/template/916416172562358536

And, oh yeah, a shit ton of video editing and Photoshopping. I think I've just developed an addiction to making AI porn lol.

 

[AmateurFaker]

You can use the loras in civit ai if you upload them publish them in tensor. Here is how.
Download lora's from here:

Civitai | Share your models

civitai.com

OR HERE:

CivArchive (CivitAI Archive) - AI Model Archive

Browse and discover AI models from CivitAI, TensorArt, and SeaArt. Community-driven archive preserving access to AI resources.

civitaiarchive.com

1- Click on Host my Model.

2- Follow image

3-Click on host a model

4-Upload your safetensor file here. You can upload a dummy jpg file for showcase

5- Wait for it to activate and copy the code and paste it into a template.


Active templates:
Pro: https://tensor.art/template/891334572870999426
Pro: https://tensor.art/template/876745929397411027
Free 3 sec only: https://tensor.art/template/881072997278956171