Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

video-production

Étoiles2

Forks0

Mis à jour6 avril 2026 à 14:00

Video production best practices and learnings for AI-generated content. Triggers on "video quality", "production tips", "Seedance tips", "Kling tips", "lip-sync", "ffmpeg composition", or video generation troubleshooting.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

albertnahas

albertnahas/content-factory

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Techniciens audio et vidéoArts, design, divertissement, sports et médias·SOC 27-4011

SKILL.md

readonly

name	video-production
description	Video production best practices and learnings for AI-generated content. Triggers on "video quality", "production tips", "Seedance tips", "Kling tips", "lip-sync", "ffmpeg composition", or video generation troubleshooting.

Video Production Best Practices

AI Video Generation

Seedance (Image-to-Video)

Use Pro 480p without audio for cost efficiency (best price/quality ratio)
Camera presets are in services/presets/camera.ts — use them instead of writing custom prompts
For loops: use lastFrameImage parameter (Pro only) with the same image as start
Duration sweet spot: 5 seconds (longer = more cost, diminishing returns)

Kling V2.6 (UGC Ads)

Silent mode (generate_audio: false) DOES produce lip movements when prompt includes dialogue
Always include "She says: ..." or "He says: ..." in the prompt
Keep prompts natural and simple — over-describing produces fake results
Always review silent video before spending on lip-sync

p-video (Audio-Driven Animation)

Best for character animation (talking food, objects, mascots)
Resize input images to exactly 720x1280 for 9:16
Takes 2-5 minutes per shot

PixVerse Lipsync

Best results of any lip-sync model
Preserves resolution, FPS, and audio quality
Do NOT use Sync Lipsync 2 Pro or LatentSync

Post-Production with ffmpeg

Speed Adjustment

When creating 1.2x speed versions, voiceover MUST also be sped up (atempo=1.2)
Background music stays at normal speed
Captions timestamps must be divided by speed factor

Audio Mixing Levels

Voiceover: 100% (primary)
Background music: 12-15% (subtle bed)
SFX (whoosh, impacts): 40-50% (noticeable but not overpowering)
Always use normalize=0 in final mix to prevent clipping

ASS Captions

Generate from Whisper word-level timestamps
Scale timestamps by speedup factor when video is sped up
Bold white text with black outline for readability

Expression Guidelines

Expression = Enthusiasm, NOT Aggressiveness
AI avatars must appear friendly, warm, enthusiastic
"Excited friend sharing a discovery" > "angry person lecturing"
Avoid angry, frustrated, or confrontational expressions

Video Reversal Technique

Generate 10s of AI video → get 20s of content via reversal
Use ffmpeg: ffmpeg -i input.mp4 -vf reverse -af areverse reversed.mp4
Great for disassembly/assembly, transformation, or before/after content

Plus depuis ce dépôt

même dépôt

content-repurposing

albertnahas/content-factory

Content repurposing strategy and extraction rules — how to derive maximum formats from a single generated asset. Triggers on "repurpose", "reuse content", "turn this into", "make a carousel from this video", or when /repurpose command is invoked.

2026-04-072

marketing-plan

albertnahas/content-factory

Create a full marketing plan for any project — market research, positioning, funnel strategy, content roadmap, channel mix, and KPIs. Triggers on "marketing plan", "marketing strategy", "go-to-market", "GTM plan", "marketing roadmap", "how to market this", or when a user needs a comprehensive marketing blueprint for a product or service.

2026-04-072

ad-script-writing

albertnahas/content-factory

Proven ad script frameworks and copywriting patterns for video ads. Triggers on "write ad script", "ad copy", "hook ideas", "curiosity loop", "script structure", or when any /generate-ad or /generate-ugc-ad command is invoked.

2026-04-062

content-strategy

albertnahas/content-factory

Content strategy guidance for social media marketing — category psychology, posting cadence, retention optimization, and content planning. Triggers on "content strategy", "posting schedule", "what to post", "content calendar", "category balance", or engagement optimization questions.

2026-04-062

gen-ai-media

albertnahas/content-factory

AI model selection guidance for media generation — which image, video, TTS, and music model to use for each use case. Triggers on "which model", "best model for", "image generation model", "video model", "TTS provider", or cost optimization questions.

2026-04-062

name	video-production
description	Video production best practices and learnings for AI-generated content. Triggers on "video quality", "production tips", "Seedance tips", "Kling tips", "lip-sync", "ffmpeg composition", or video generation troubleshooting.

Video Production Best Practices

AI Video Generation

Seedance (Image-to-Video)

Use Pro 480p without audio for cost efficiency (best price/quality ratio)
Camera presets are in services/presets/camera.ts — use them instead of writing custom prompts
For loops: use lastFrameImage parameter (Pro only) with the same image as start
Duration sweet spot: 5 seconds (longer = more cost, diminishing returns)

Kling V2.6 (UGC Ads)

Silent mode (generate_audio: false) DOES produce lip movements when prompt includes dialogue
Always include "She says: ..." or "He says: ..." in the prompt
Keep prompts natural and simple — over-describing produces fake results
Always review silent video before spending on lip-sync

p-video (Audio-Driven Animation)

Best for character animation (talking food, objects, mascots)
Resize input images to exactly 720x1280 for 9:16
Takes 2-5 minutes per shot

PixVerse Lipsync

Best results of any lip-sync model
Preserves resolution, FPS, and audio quality
Do NOT use Sync Lipsync 2 Pro or LatentSync

Post-Production with ffmpeg

Speed Adjustment

When creating 1.2x speed versions, voiceover MUST also be sped up (atempo=1.2)
Background music stays at normal speed
Captions timestamps must be divided by speed factor

Audio Mixing Levels

Voiceover: 100% (primary)
Background music: 12-15% (subtle bed)
SFX (whoosh, impacts): 40-50% (noticeable but not overpowering)
Always use normalize=0 in final mix to prevent clipping

ASS Captions

Generate from Whisper word-level timestamps
Scale timestamps by speedup factor when video is sped up
Bold white text with black outline for readability

Expression Guidelines

Expression = Enthusiasm, NOT Aggressiveness
AI avatars must appear friendly, warm, enthusiastic
"Excited friend sharing a discovery" > "angry person lecturing"
Avoid angry, frustrated, or confrontational expressions

Video Reversal Technique

Generate 10s of AI video → get 20s of content via reversal
Use ffmpeg: ffmpeg -i input.mp4 -vf reverse -af areverse reversed.mp4
Great for disassembly/assembly, transformation, or before/after content