Full-stack AI video creation pipeline. Supports Sora/Kling/Runway/Pika/即梦/Luma/Vidu/海螺AI. From concept→storyboard→frame-level prompts→platform optimization→p...
--- name: AI Video Director slug: ai-video-director description: Full-stack AI video creation pipeline. Supports Sora/Kling/Runway/Pika/即梦/Luma/Vidu/海螺AI. From concept→storyboard→frame-level prompts→platform optimization→post-production. 100+ cinematic shot templates, 10+ video types (ads/shorts/microfilms/animations). author: Marvis Studio version: 1.0.0 tags: ["ai-video", "video-generation", "sora", "kling", "runway", "prompt-engineering", "storyboard", "content-creation"] requires: [] metadata: emoji: 🎬 category: "AI/Content Creation" pricing: "Free" --- # AI Video Director ## Capabilities | Capability | Description | |------------|-------------| | **Multi-platform support** | Sora, Kling, 可灵, Runway Gen-4, Pika 2.0, 即梦, Luma Dream Machine, Vidu, 海螺AI, PixVerse | | **Video type templates** | 10 types: Ad TVC, Brand promo, Short video, Microfilm, Product showcase, Educational, MG animation, Holiday marketing, Plot twist, Visual masterpiece | | **Professional storyboarding** | Shot-by-shot breakdown: shot number, scene, camera movement, duration, visual description, prompt, sound effect, voiceover | | **Frame-level prompt precision** | 6-dimensional description: subject action, lighting, color, material, atmosphere, composition | | **Platform parameter optimization** | Resolution, frame rate, duration, aspect ratio, motion intensity, consistency parameters | | **Post-production guidance** | Editing rhythm, transition recommendations, BGM suggestions, subtitle styling | | **Multi-version output** | Standard, Creative, Minimalist versions | ## Trigger Keywords - "AI视频", "video制作", "分镜头", "storyboard", "AI动画", "Sora", "Kling", "Runway", "视频提示词", "video creation" ## Workflow ### Phase 1: Requirement Analysis Identify video type, target platform, style preference, duration, brand tone. ### Phase 2: Creative Strategy Develop narrative structure (flashback/parallel/progressive/suspense), determine visual style reference. ### Phase 3: Storyboard Generation Output professional shot-by-shot storyboard table with 6D prompts and platform-specific parameters. ### Phase 4: Post-production Plan Output editing rhythm suggestions, transition recommendations, BGM style recommendations, subtitle styling. ## Output Format - **Storyboard Table** (Markdown table with columns: Shot#, Scene, Camera, Duration, Visual Description, Prompt, Sound, Voiceover) - **Platform-specific Prompt Sets** (formatted for each target platform) - **Parameter Recommendations** (JSON with resolution, fps, duration, etc.) - **Post-production Checklist** ## Examples **Example 1: Product Showcase Video** ``` Input: "Create a 30-second product showcase for a new smartphone" Output: - Storyboard: 8 shots (intro→features→demo→call-to-action) - Sora prompts (English, detailed frame descriptions) - Platform: Sora, 1920x1080, 30fps, 30s - BGM: Upbeat electronic ``` **Example 2: Brand Story Microfilm** ``` Input: "Tell a brand story about sustainability in 60 seconds" Output: - Narrative: Problem→Solution→Impact - 12-shot emotional arc - Kling prompts (Chinese preferred) - Color grading: Natural, earthy tones ``` ## Notes - Different platforms have different prompt formats; automatically adapt - Sora and Kling prompts work best in English - Video duration should match platform characteristics (TikTok: 15-60s, YouTube: 30s-3min) - Complex scenes should have at least 8 shots for proper storytelling - Always include negative prompts to avoid unwanted elements ## Installation ```bash npx clawhub install ai-video-director ``` ## License MIT-0
don't have the plugin yet? install it then click "run inline in claude" again.