This project addresses the creation of dynamic short videos using AI, building upon the MoneyPrinterTurbo framework. It integrates Alibaba's Wanxiang Wan2.1 for text-to-video and image-to-video generation, offering users a more controllable and AI-driven workflow for producing video content. The target audience includes content creators and developers seeking automated video production tools with flexible AI asset integration.

How It Works

The system leverages large language models (LLMs) to generate video scripts, storyboards, and prompts. These prompts then feed into Alibaba's Wanxiang Wan2.1 model to create video material through text-to-video or image-to-video processes. The generated clips are automatically stitched, edited, and combined with subtitles, voiceovers, and music to form a complete short video. This approach allows for AI-driven content ideation and visual asset generation, offering users granular control over the final output.

Quick Start & Requirements

• Primary install / run command: Deployment follows the MoneyPrinterTurbo pattern. For Windows, replace the MoneyPrinterTurbo folder with MoneyPrinterAICreate and update the start.bat script. A Baidu Netdisk link is provided for deployment.
• Non-default prerequisites and dependencies: Requires a Wan2.1 API key, obtainable for free from the "魔搭社区" (Modao Community). A VPN may be necessary for edge-tts functionality if connection issues arise.
• Estimated setup time or resource footprint: Video material generation via the Wan2.1 trial API is notably slow, with each 5-second clip potentially taking 10-20 minutes or more.
• Links: Wan-2.1 trial page, Wan-2.1 API key application, Wan-2.1 project repository.

Highlighted Details

• Integrates Alibaba's Wanxiang Wan2.1 for AI-generated video materials, supporting both text-to-video and image-to-video.
• Features AI-driven generation of video scripts and storyboards, with options for manual prompt modification.
• Automates the stitching and editing process to produce complete short videos, including subtitles, voiceovers, and background music.

Maintenance & Community

Future optimizations are planned, including integration with Tencent Hunyuan LLM, prompt quality enhancements, and revised editing logic. No specific community channels (e.g., Discord, Slack) or notable contributors are detailed in the provided README.

Licensing & Compatibility

The specific open-source license is not stated in the provided text. Compatibility for commercial use or integration with closed-source projects is also not specified.

Limitations & Caveats

The project relies on the Wanxiang Wan2.1 trial API, which exhibits slow generation speeds (10-20+ minutes per 5s clip) and occasional failures. Users should be aware of privacy implications, as images uploaded for Wan2.1 generation are collected. Prompt generation for Wan2.1 has limitations regarding special characters. LLM outputs can sometimes be empty or erroneous, requiring retries.