Prompts de MiniMax H3 con ejemplos de vídeo reales

Kōda · @aimikoda

A ten-shot character reveal whose contour lines, graphic accents and music adapt to the supplied design.

ComfyUI · @ComfyUI

A basic motion blockout and reference images guide a polished vertical shot, with an optional second pass upscaling to 2K.

不満福 · @fumanfuku

A reference walking video transfers runway motion to a consistent stylized character in a 15-second vertical R2V result.

知识猫AI实验室 · @GeekCatX

A video reference is converted into seamless matte-gray clay models while preserving every subject, action, beat and camera move.

toba · @999toba

Two interpretations of the same image concept are animated with MiniMax H3 to compare how upstream visual design changes the final motion.

Kent Dhani · @kentdhani

A reference-motion experiment shows strong movement transfer while openly exposing remaining face and photorealism weaknesses.

モジョン · @mojon1

A static Blender render is animated with MiniMax H3, demonstrating how conventional 3D design can anchor generative motion.

GENFILM LAB · @Claud31150Code

Fictional printed photographs become moving memories, with period-specific textures and music-aligned transitions.

toyxyz · @toyxyz3

A ComfyUI test explores an Anime-to-Realism reference-video LoRA for H3.

⚡️Mirthtime ⚡️ · @mirthtime

A creator tests facial expressions using text prompts alone with H3.

akatz · @akatz_ai

The adapter creator tests a 1,000-step character-swap LoRA, preserving source motion and audio while documenting drift.

SmallKino · @SmallKino_

A comparison changes reference-video size to explore its influence on VFX transfer and character output.

SmallKino · @SmallKino_

A side-by-side character-swap test compares LoRA strength 1.0 on the left against native H3 on the right.

烁皓 · @eternityspring

A dance video and full-body photo drive character replacement, with longer videos generated in segments.

元荒川ちゃん❤ · @mtArakawa

A resin-encapsulation concept combines NBPro imagery with MiniMax H3 animation.

serva huang · @servasyy

A high-reach comparison places original footage on the left and a MiniMax H3 recreation on the right.

CALMDUST aka ちゃんこあ · @core_tan

A WebUI example stacks character images into one video reference and combines character and audio RefMod with FlashGen.

Mike Pevzner · @Mike_Pevzner

A creator shares a family-involved Recast demonstration, linked to fal’s H3 Max character replacement announcement.

kuripon · @malonpon

H3 Max uses an AI-generated 3D hand-dance previs and character sheet to test face fidelity at awkward camera angles.

The Rogue AI · @TheRogueA1

A creator uses MiniMax H3 inpainting to insert themself into a recognizable Ripley scene from Alien.

Kent Dhani · @kentdhani

A local Ref2V tracking shot keeps one airport arrival in focus amid bodyguards, fans, paparazzi and flashing cameras.

Jonathan Fischoff · @jfischoff

An H3 Max Insert Video demonstration presents the author’s video-editing result in a longer source clip.

toyxyz · @toyxyz3

A high-reach ComfyUI H3 character-performance test uses Malfoy imagery in a 20-second source compilation.

ひねもす · @hinemos99667732

A fan-art video combines a community H3 Turbo variant and ControlNet with image generation, Blender, Krita and upscaling tools.

Modengsir AI · @ModengSir

A short H3 test checks whether spoken counting and raised fingers remain logically aligned.

adrián · @myothercoolname

An author demonstrates a 0.928-second H3 Max reference-to-video clip using fal demo characters.

Rehan Sheikh · @rehan_shei

An H3/Max equirectangular LoRA is demonstrated in a VR headset, extending panoramic generation into immersive viewing.

ささゆう｜MiniMax H3・AI撮影スタジオKonte開発 · @forticodesign

A creator reports stronger reproduction when specifying the 3D underlay data in an H3 generation.

ヒロ(HG)｜実務で使えるAI活用 · @hsrk_g_hsrk

A fan short combines canonical character sheets and MCP character context with H3 video, Irodori TTS and Remotion.

さわ|AI画像生成で月3桁継続中 · @AI_swwww

Three identity images guide six timed hotel-room selfie moments with subtle acting, handheld shake and imperfect front-camera texture.