xhongc/ai_story

What it solves

AI Story automates the complex process of turning a simple story theme or outline into a professional video. It eliminates the need for manual scriptwriting, storyboard design, image generation, and video editing by orchestrating multiple AI models into a single pipeline.

How it works

The platform uses a modular pipeline (Responsibility Chain pattern) to move a project through several automated stages:

1. Scripting: AI rewrites a theme into a full video script.
2. Storyboarding: The script is split into scenes with descriptions and prompts.
3. Image Generation: AI creates high-quality images based on the storyboard prompts.
4. Camera Planning: AI determines the best camera movements (zoom, pan, etc.) for each scene.
5. Video Production: Static images and camera parameters are converted into dynamic video clips using AI video tools.

Who it’s for

• Short-video creators making content for platforms like TikTok or Douyin.
• Educators turning knowledge-based text into visual explanations.
• Parents creating illustrated children's stories or bedtime tales.
• Creative artists prototyping video scripts and storyboards quickly.

Highlights

• End-to-end automation: One-click generation from theme to final video.
• Multi-model support: Integrates with various LLMs (GPT, Claude) and image/video generators (Stable Diffusion, Midjourney, Runway, Pika).
• Granular control: Allows manual adjustment of scripts, prompts, and camera movements at any stage.
• Robust infrastructure: Uses Celery for asynchronous task processing and Docker for easy deployment.
• Load balancing: Intelligent scheduling across multiple AI service providers to ensure stability.