aaronyi97/image-story-video-wizard

What it solves

This project provides a structured, step-by-step workflow for creating AI-generated story videos. It prevents the overwhelming feeling of managing a complex production pipeline by breaking the process down into manageable stages, ensuring that critical decisions (like visual style and script finalization) are confirmed before moving to expensive or time-consuming generation tasks.

How it works

The tool acts as a "Skill" for Codex and WorkBuddy, functioning as a state machine that guides the user through 16 distinct production stages—from initial topic selection and benchmarking, through scriptwriting, voice-over, storyboarding, and image generation, to final rendering. It uses a PROJECT_STATE.json file to track progress, allowing users to switch between different AI assistants or tools while maintaining the project state.

Who it’s for

Content creators making AI-narrated books, historical or emotional stories, and podcast-style videos that require consistent characters, a fixed visual style, and static image storytelling.

Highlights

• Guided Workflow: Every interaction clearly states the current stage, what the tool is doing, what the user needs to provide, and what the next step is.
• State Persistence: Project progress is saved in a JSON file, allowing for seamless transitions between different AI hosts.
• Quality Control: Built-in checkpoints ensure that scripts and visual styles are confirmed before bulk image generation or final rendering occurs.
• Tool Agnostic: Integrates with external capabilities like Seed-TTS 2.0 for voice and HyperFrames for assembly, while verifying tool availability before proceeding.