Courses

Video Generation


Video Hailuo Minmax Higgsfield
Runway
Open-source Video Models

Remove background
Face swap
character swap
video -> animation

https://huggingface.co/tencent/spaces

Adobe - Firefly, Creative Suite

VIDEO GENERATORS

https://dreamina.capcut.com/ - free only, no unlimited
* Leonardo - $60/month, unlimited videos at slower pace, with multiple submissions, choose video models

* MidJourney - $60/month, unlimited videos at slower pace, with multiple submissions, choose video models

* https://wavespeed.ai/

freepik

Hailuo MiniMax - $35/month, 100 videos

Kling 2.1 - Kuaishou

Krea AI $60/month, choose video models

RunwayML - @runwayml, upload our video in Gen3, and select Restyle Video. turn video into animation

HeyGen - $29/month, avatar, unlimited videos, Voice Mirroring emotion, languages

Luma Labs - $95/month, unlimited videos, Ray2 Flash, 3x faster, 3x cheaper new model.

Vidu - $99/mo unlimited videos

Nex Art - brand marketing, tools

https://fal.ai/pricing

https://seedanceai.net/pricing

OPEN-SOURCE VIDEO

gradio
comfyui

https://github.com/chengzeyi/Comfy-WaveSpeed

Wan video gen

https://huggingface.co/Wan-AI

Wan 2.1, Framepack - Alibaba

https://github.com/lllyasviel/FramePack

Tencent Hunyuan - 3D generation model. It comes with two versions: 3D 2.0 MV (Multi-View Generation) and 3D 2.0 Mini for faster generations.

ByteDance - Seedance

https://github.com/Lightricks/LTX-Video
https://huggingface.co/Lightricks/LTX-Video

https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/ltxv-13b-i2v-long-multi-prompt.json Alibaba - EMO

https://github.com/ali-vilab/VACE

https://github.com/Saiyan-World/goku (Bytedance, Goku)
https://huggingface.co/datasets/saiyan-world/Goku-MovieGenBench

https://github.com/ai-forever/KandinskyVideo

https://github.com/kijai/ComfyUI-LivePortraitKJ

https://github.com/Doubiiu/ToonCrafter

Descript - Unique for editing videos by editing the script, ideal for podcasters and content creators, with AI-driven features for transcription and voice synthesis.

RE-USE

1.Re-mixing TikTok viral videos in Chinese
Re-mixing Douyin in Chinese

Motion Reference (Video) - guides movement
Character Reference (foreground)
Background Reference
1. Video Background Removal - Separate the main character.
2. Create Video Backgrounds

3. Captioning

{
shot: {
composition: "Medium tracking shot, 50mm lens, shot on RED V-Raptor 8K with Netflix-approved HDR setup, shallow depth of field",
camera_motion: "smooth Steadicam walk-along, slight handheld bounce for naturalistic rhythm",
frame_rate: "24fps",
film_grain: "clean digital with film-emulated LUT for warmth and vibrancy"
},
subject: {
description: "A young woman with a petite frame and soft porcelain complexion. She has oversized, almond-shaped eyes with long lashes, subtle pink-tinted cheeks, and a heart-shaped face. Her inky-black bob is slightly tousled and clipped to one side with a small red strawberry hairpin. Her style blends playful retro and modern Tokyo streetwear: she wears a crocheted ivory halter top with scalloped edges, high-waisted denim shorts with a wide brown belt and a red enamel star buckle, and a loose red gingham blouse draped off one shoulder. Her accessories include glossy cherry lip tint, a beaded bracelet stack, and soft shimmer eyeshadow.",
wardrobe: "Crocheted ivory halter with scalloped trim, fitted high-waisted denim shorts, wide tan belt with red enamel star buckle, oversized red gingham blouse slipped off one shoulder, strawberry hairpin in side-parted bob, and translucent plastic bead bracelets in pink and cream tones."
},
scene: {
location: "a quiet urban street bathed in early morning sunlight",
time_of_day: "early morning",
environment: "empty sidewalks, golden sunlight reflecting off puddles and windows, occasional birds fluttering by, street slightly wet from overnight rain"
},
visual_details: {
action: "she walks rhythmically down the sidewalk, swinging her hips slightly with the beat, one hand gesturing playfully, the other adjusting her shirt sleeve as she sings",
props: "morning mist, traffic light turning green in the distance, reflective puddles, subtle sun flare"
},
cinematography: {
lighting: "natural golden-hour lighting with soft HDR bounce, gentle lens flare through morning haze",
tone: "playful, stylish, vibrant",
notes: "STRICTLY NO on-screen subtitles, lyrics, captions, or text overlays. Final render must be clean visual-only."
},
audio: {
ambient: "city birds chirping, distant traffic hum, her boots tapping pavement",
voice: {
tone: "light, teasing, and melodic",
style: "pop-rap delivery in Japanese with flirtatious rhythm, confident breath control, playful pacing and bounce"
},
lyrics: "ラーメンはもういらない、キャビアだけでいいの。 ファイナンスのおかげで、私、星みたいに輝いてる。"
},
color_palette: "sun-warmed pastels with vibrant reds and denim blues, soft contrast with warm film LUT",
dialogue: {
character: "Woman (singing in Japanese)",
line: "ラーメンはもういらない、キャビアだけでいいの。 ファイナンスのおかげで、私、星みたいに輝いてる。",
subtitles: false
},
visual_rules: {
prohibited_elements: [
subtitles,
captions,
karaoke-style lyrics,
text overlays,
lower thirds,
any written language appearing on screen
]
}
}


Prompt Description

Composition type (medium tracking shot, 50mm lens)
Motion style (Steadicam, with a touch of handheld)
Frame rate and LUT film grain
You basically get full cinematographer-level control here.
Subject & Wardrobe described in visual, tactile language

Scene & Environment
Time of day: Early morning
Atmosphere: Golden light, empty street, wet pavement
It even includes birds and puddle reflections.
Visual Details & Props
Physical actions like walking, singing, adjusting clothes
Elements like sun flares and mist
Props (traffic light in distance, puddles, etc.)
Lighting & Tone
Golden hour with HDR bounce and soft lens flares. Think soft, dreamy, but vibrant. It also sets the mood: “playful, stylish, vibrant.”

Audio & Lyrics
Ambient audio: birds, distant cars, shoes tapping
Voice tone: melodic, teasing, playful
Lyrics in Japanese: flashy, finance-themed
No subtitles, no captions—this is a strict “visual-only” policy.

You can plug in your own style references, film gear, mood, and tone. The more specific, the better.

Tips to Nail the Perfect Veo JSON Prompt
Stick to film language: Use words like “lens,” “frame rate,” “cinematic motion,” “bokeh,” etc.
Describe subject like you’re painting: Facial structure, clothing texture, accessories
Set tone with lighting and audio: Warm/cold, sharp/soft, ambient/clean
Use verbs: Have your character walk, spin, sing, adjust, etc.
Avoid prohibited elements: Like this JSON did—no on-screen text unless you want chaos.

17network.com