A Flutter-based mobile application, DirectorAI, enables users to generate comic drama videos from text prompts. It automates the creation of scripts, storyboards (images and short videos), and final video synthesis, targeting creators seeking efficient mobile video production.

How It Works

DirectorAI employs an AI agent architecture using the ReAct (Reasoning + Acting) loop. A core LLM (GLM-4.7) interprets user input, plans actions, and orchestrates calls to specialized AI models. It integrates Google's Gemini for text-to-image generation and Veo for image-to-video synthesis, composing these elements into a complete comic drama video from a single text description.

Quick Start & Requirements

• Installation: Clone the repository, run flutter pub get.
• Prerequisites: Flutter SDK >= 3.0.0, Dart SDK >= 3.0.0, Android Studio/VS Code, Android device/emulator.
• Configuration: Requires API tokens for Zhipu AI (GLM-4.7) and potentially Google Cloud services (Gemini, Veo). API tokens must be configured securely (e.g., via environment variables or secure storage, not committed to Git).
• Running: Execute flutter run.
• Links: No direct links to official quick-start guides or demos are provided in the README.

Highlighted Details

• AI Orchestration: Utilizes a ReAct agent loop for intelligent task planning and execution.
• Multi-Model Integration: Seamlessly integrates GLM-4.7 (decision-making), Gemini (image generation), and Veo (video generation).
• Mobile-First: Developed using Flutter for cross-platform mobile deployment.
• Customization: Offers configurable parameters for character consistency and scene planning.

Maintenance & Community

The project is maintained by freestylefly. A WeChat public account ("苍何") is mentioned for community discussion, but no direct links to Discord, Slack, or a public roadmap are provided.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The MIT license is permissive and generally compatible with commercial use and closed-source linking.

Limitations & Caveats

The application's functionality is heavily dependent on the availability and performance of external AI model APIs, requiring valid API keys and potentially incurring usage costs. Security best practices for API token management are crucial and must be implemented by the user. The README does not detail performance benchmarks or specific limitations of the AI models themselves.