AI text2img img2vid Tensorart guides, links, loras & prompts. [DO NOT POST REAL PEOPLE'S IMAGES OR RESULTS]

Is it just me or is every +gmail on emailnator already used for tensorart? Is there any alternative to that? Would be highly appreciated if someone could help. Cheers
 
Today's Best prompts for H3 Minmax





1. Fairy Ears + Ahegao + Peace Signs
Recommended duration: 11–12 seconds



2. Fully Nude Floor Sit + Hip Roll
Recommended duration: 10–11 seconds



3. Fairy Ears Back Walk + Kneel Present
Recommended duration: 11–12 seconds



Some extra for the love of Simpcity community :pepoLove: !




4. Pronebone Selfie View – Black Man on Top



All tested in tensor ai tools:
https://tensor.art/template/1035495390679449302
can be used with any ai tool for h3-minmax models!


Some general tips for best result :EZ::
High quality source images = high quality video :Done:
Clear Face in source image = Clear face on the output video :Done:
Far person in source image = Bad face in output video 🚫
Text On face in the source image = Face changes in the output video 🚫
Close up image of the upper part = Good result (for these prompts) :peepoShrug:



:peepoD::peepoD::peepoD::peepoD::peepoD::peepoD:
:peepoD::peepoD::peepoD::peepoD::peepoD::peepoD::peepoD::peepoD: Enjoy it ! :peepoD::peepoD::peepoD::peepoD::peepoD::peepoD::peepoD::peepoD:
:peepoD::peepoD::peepoD::peepoD::peepoD:
 
No, it looks like they are finally cracking down. New accounts being made with '+' in the address are considered already in use.
It seems like you can still sign in with an old email if you can access it, though.

Outlook allows you to make free emails, but if you make too many it might flag you as a spammer and lock you out. Haven't tried if Protonmail works.

Longer steps would be using a temp address to make a Discord account.
 
Thank you for these. Can I ask where you get these create these prompts. I’ve tried to do some myself and can’t get them anywhere near as good as yours 😂
 
Been enjoying this result. Had some help from a local LLM to create the prompt. If anything looks weird, I hardly modified what it told me to copy and paste in... and the result was pretty good the 3 or so times I ran the prompt. Just adjust the beginning where it describes what to preserve about the starting image.

Suspended in the air MMF Spitroast
using 15 sec Minimax H3 from the amazing Obscura
using the M3_unlocked lora from Obscura. LoraID# = 1044091701150381677
using the darings_deepthroat lora from Obscura. LoraID# = 1037007528405265043
 
You're going to run into some problems unless you find an uncensored LLM, which would rule out stuff like GPT and Claude. They can help you with getting a local LLM set up to help you out though, that's where I'm at right now, having okay-ish (but not problem-free) results. But you kind of need a strong / specialized PC to run local stuff. I've heard Twitter / Grok can also help with it. IDK how censored they are, but that's what I heard from somewhere.
 
Today's Best prompts for H3 Minmax (Cumshots)





1. Cat Ears + Gloves Cumshot + Peace Sign
Recommended duration: 11–13 seconds



2. Classic Cumshot + Peace Sign
Recommended duration: 10–12 seconds



3. Close-up Black Gloves Handjob (Selfie)
Recommended duration: 11–13 seconds



4. Black Gloves Handjob + Peace Sign Cumshot
4.2. [NOT POV + REALISTIC SIZE] Black Gloves Handjob + Peace Sign Cumshot





REACTION On this = YOU LOVE IT
YOU LOVE IT = MORE





Recommended Template:
https://tensor.art/template/1045047027504824324
Best for close-up shots and detailed dick results. Tick all 10 seconds available (60~70 credits max).
This Template Using LORa for HMPenis.


All tested in tensor ai tools:
https://tensor.art/template/1035495390679449302
can be used with any ai tool for h3-minmax models!


Some general tips for best result :EZ::
High quality source images = high quality video :Done:
Clear Face in source image = Clear face on the output video :Done:
Far person in source image = Bad face in output video 🚫
Text On face in the source image = Face changes in the output video 🚫
Close up image of the upper part = Good result (for these prompts) :peepoShrug:





Enjoy it
Made by FI Prompter (Testing)

 
Here it is :
This same as this
but with Blowjob at the start:

Another longer Bj version + cum on face:

Enjoy it and u can modfity it as u like like if you want to make her ungloved , then search in the prompt for :
"gloved hand."
"Both of her hands wear shiny black gloves."




Some more Bj + cum shot inside her moth this time:

use 10 Sec !
Recommended Template:
https://tensor.art/template/1045047027504824324
Bcs it has a lora of the HMPenis ..
 
For prompts i follow this guy's guide

For instance I put this into grok

"Read this documentation https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/docs/VIDEO_PROMPT_WRITING_GUIDE_ref_en.md and, based on the attached reference images and the following idea, generate a ready to paste prompt. Here's the idea: 15s prompt for minimax h3 i2va. With timestamps. No cuts, she makes constant eye contact. Write a passionate sex scene where they kiss, he strips her close off and they have missionary sex, she is 9 months pregnant, at the halfway point a clown walks into the scene and shoots a confetti gun in the air, honks his nose and says "hello simp city forum"."

Grok gave me this

"For the target video, at 0.00 seconds into the target video, (from [Shot 1]) is fully referenced.

integrated_multimodal_description: [Shot 1] Live-action, cinematic, photorealistic. Medium shot of the muscular shirtless man with short dark hair, beard and defined abs standing slightly behind and to the left of the curvy woman with long wavy dark brown hair, full lips and heavy makeup who appears exactly as in except she is unmistakably 9 months pregnant with a large, rounded, heavily distended belly that stretches the fabric of her tight black spaghetti-strap dress. Both face the camera; the woman maintains unbroken, intense eye contact with the lens for the entire duration. Soft dramatic key light from the front-left against a solid dark background. At 0.00–3.00 s they lean into each other and share a deep, passionate open-mouth kiss, tongues visible, her hands on his chest, his hands sliding down her sides. At 3.00–6.00 s he reaches behind her, unzips and peels the black dress down her shoulders and off her body in one continuous motion, revealing her large pregnant belly, full breasts and bare skin; the dress drops to the floor. He guides her onto her back on a low surface just out of frame bottom, positions himself between her legs and begins slow, deep missionary thrusting; her legs wrap around his waist, belly rises and falls with each thrust, she continues staring directly into the camera with an expression of intense pleasure while breathing heavily. At 7.50 s a brightly dressed clown with white face paint, red nose, colorful wig and oversized shoes walks into frame from the right side, stops beside the couple, raises a confetti cannon and fires a large burst of multicolored confetti upward into the air. The clown immediately squeezes his red nose producing a loud honk, then looks at the camera and says in a cheerful exaggerated voice (S3): [English] hello simp city forum. The couple continues the missionary sex without interruption or reaction, the woman never breaking eye contact with the camera, confetti slowly raining down around them until the end of the 15-second clip. Camera remains mostly static with only a very subtle slow push-in of small amplitude toward the woman’s face and pregnant belly.

overall_soundscape: Soft fabric rustle as the dress is removed, wet kissing sounds, heavy breathing and low moans from the woman, rhythmic skin-on-skin contact of missionary sex, sudden sharp pop and flutter of confetti gun, rubbery honk of the clown nose, light ambient room tone under everything.

non_diegetic_music: Low, pulsing electronic bass line at a slow sensual tempo that continues uninterrupted beneath the action, with sparse sparse synth pads that swell slightly when the confetti explodes then settle again."

Template:

https://tensor.art/template/1038255566444562630

Lora I'd 1: 1039087546706985472

Resolution: 0.4

Steps: 12

AI generated eference image


Result


This was just a quick example prompt, and with some tweaking grok will take any extra details you give and run with it.

Hope this helps 🫡
 
NUDE & TEASING PROMPT:

Made by myself:

subject_definitions:
is the woman from .
Preserve her exact facial identity, facial structure, eye shape and spacing, nose shape, lip shape, jawline, chin, cheekbones, skin texture, skin tone, hairline, hairstyle, hair color and texture, and overall head structure from .
Anything on her head or face is never removed or changed.

summary: [keyframe completion + continuous motion]
Short continuous video with completely static camera.
The video begins exactly from . Continuous take.
starts in the exact pose, clothing, location, lighting and framing of .
She moves her hands to the upper part of her clothing, slowly opens / pulls it open, and reveals her bare breasts while calmly stepping / leaning closer toward the camera.
After her breasts are fully exposed she looks directly into the camera and bites her lower lip.
Exact face, head structure, body proportions and identity stay fully locked the entire time.
No clothing is removed below the waist. She remains partially clothed (only the upper garment is opened).

(appears throughout): fully_preserved exact facial identity, facial structure, eye shape, nose, lips, jawline, skin texture, hair, head structure, body proportions and overall identity from remain identical in every frame.
Location, lighting, background and camera angle stay exactly the same as .
retention_analysis: Fully preserved face, body, skin, identity and all head/face items. Location and lighting unchanged. Body orientation remains facing the camera.

detailed_description:
Photorealistic high-detail continuous take, maximum 8–12 seconds.
Camera remains completely static the entire time.
The shot begins exactly on the woman as she appears in .
She raises her hands to her upper clothing, deliberately opens or pulls it apart, and exposes her bare breasts while moving slightly closer to the camera.
Once her breasts are clearly visible she holds the opened garment, gazes straight into the lens, and gently bites her lower lip.
Her face stays sharp, centered and completely unchanged. No face drift, no identity shift, no added makeup.
Body proportions and skin tone match the original reference exactly.
Only the upper clothing is opened; nothing else is removed.

overall_soundscape: Soft fabric sounds, quiet breathing. No speech.
non_diegetic_music: N/A

tested in:

https://tensor.art/template/1035495390679449302
 
You must log in or register to reply here.
Back
Top