Multi-segment missionary scenario
Details
Download Files
Model description
How to Use
This is a collection of multiple workflows for a complete I2V missionary scenario. I did not bother to include a sample here, but it works wonders with two subjects as well.
Contrary to the deepthroat scenario that I posted earlier this workflow set relies on scene cuts. It is enabled by this LORA that can take any full-body or portrait picture of subject(s) in ANY state in ANY setting and generate missionary insertion with these subject(s) in ANY state in ANY setting. The flexibility that it allows is mind-blowing. I will leave the rest to your imagination and experimentation. The only caveat is it will highly likely flash the original image in the first few frames even if you prompted away from it. You can either turn it into action you want or cut the frames with the new frame skip feature of workflow that allows you to skip unnecessary transitions and cut the segment to the size and actions you want
/model/1983608/wan-pov-missionary-insertion
v1.0
missionary-wan22-i2v-01-start
Actions preceding and including insertion. Uses this LORA
/model/1983608/wan-pov-missionary-insertion
missionary-wan22-i2v-02-middle
Missionary sex sequence. Uses this LORA
/model/1811313?modelVersionId=2176505
Conversely, this LORA supports many other positions
missionary-wan22-i2v-03-end-bodyshot
Ending with cum on body. Uses this LORA
/model/2031069?modelVersionId=2298673
missionary-wan22-i2v-03-end-facial
Ending with facial. I am actually not 100% sure what LORA it uses as Stability Matrix is not matching it to Civitai anymore. I want to say it is by https://civitai.com/user/LocalOptima
because of the way they upload LORAs and their naming convention, but I do not see it anymore :(
You can probably replace with other facial LORAs.
NOTE: This segment will generate a transition (from missionary into a sitting position) that is not optimal. Workflow is preset to cut that transition out at 24 frames. You might need to generate the segment without frame skip first. Then if you like it, lock the seed (copy it from file name if you did not copy prior to generation), enable frame skip and find the position at which to cut. That way it will not have to regen the whole segment.
video-merge-colormatch-upscale-interpolation
Join videos and upscale / interpolate if you have not done so during segment creation.
Includes color match (sometimes hurts more than helps, use wf below if you do not want to color match)
video-merge-upscale-interpolation
Join videos and upscale / interpolate if you have not done so during segment creation.
Tips and Tricks
You can mix and match this scenario with my multi-segment deepthroat scenario. Start from the same image. Since missionary relies on scene cuts it will fit nicely after the deepthroat.
You can use both missionary and deepthroat scenarios with two subjects (probably even more). Just make sure you describe each one in detail (position - left or right, features, clothes, body type) otherwise WAN will blend their features.
POV missionary insert LORA can generate subject sitting or lying down. It can do some other poses, but anatomy can get deformed.
The setting can be different from your original picture, for example your original picture is at a beach, but sex happens inside on a bed. In fact, if you keep the setting it will likely cause a transition instead of scene cut.
Workflow supports first and last frame. It can be very handy, for example if you do not like a segment overall but like how it ended (eyes open, no distortions or hallucinations, etc.) grab that last frame, enable last frame in the control box and use it in the new iteration to ensure it will end as you want it.
Workflow supports frame skip (from beginning or the end). It can be very handy if you want to move your subject into position but may not want / like the transition.
Workflow supports upscale and interpolation, but I would recommend disabling them while you iterate. Before you hit run, copy seed to clipboard. If you like the segment, paste seed back, enable upscale and whatever else (maybe frame skip or even interpolation) and rerun. Since generation already happened it will quickly upscale/frame skip/etc. If you forgot to copy the seed from the node, it is also part of the output filename.
