@371133686 — صانع Open Skill
@371133686
text-to-video-cinematic
Use when the user provides only a text description (no reference image) and wants to generate a cinematic, film-like video with native audio (dialogue/sound effects/music) — pure text-to-video (T2V) with no input image.
16秒文生视频
Use when the user wants a longer text-to-video clip (8s or 16s) on modest hardware — the LTX-2.0 multi-modal low-VRAM pipeline that expands a simple scene plus Chinese dialogue into a 3-shot or 5-shot English shot list and renders it at 25fps.
多角色参考生视频
Use when the user supplies two or more reference images (different characters, outfits or props) and wants all of them to appear in one video with locked identity — multi-subject reference (MSR) video generation driven by IC-LoRA guides.
图生视频
Use when the user provides a reference image (as the first frame) and wants to bring it to life as a cinematic video with native audio (dialogue/sound effects/music) — image-to-video based on the LTX-2.3 22B distilled workflow with first-frame anchoring.
minimaxh3i2v微表情
MiniMax+H3+I2V+(微表情),感觉挺好用一版本
micro-expression
Use when the user wants to generate a cinematic facial close-up micro-expression video (portrait 9:16) with a gradual emotional progression (e.g. smile → smile fading → eyes moistening → a single tear rolling down), slow, restrained, no cuts, film-grade soft lighting and realistic skin texture. A closer single-close-up shot of a person's face where emotion shifts subtly over ~10 seconds.
brand_tvc_director_2
Turns a product photo and a brand brief into a broadcast-grade TVC production package: three creative routes, a director treatment, a locked shot list, generation-ready prompts, a product-consistency protocol, sound design and platform cutdowns. Use for cinematic product films, brand commercials, hero TVCs and premium ad creative.
one_photo_product_ad
Turn a single product photo into a 10-second scroll-stopping social ad - locked hook, motion and caption, ready to post.