What does lianhuanhua-skills do?
Lianhuanhua Short Video Codex Skill: Token-saving prompt packs, external/built-in image generation, Doubao TTS, FFmpeg composition
Overview
This skill is designed for generating "Lianhuanhua" (comic strip) style short videos using AI. It integrates image generation, text-to-speech, and video composition to create engaging multimedia content efficiently.
Features
- Token-Saving Prompt Packs: Pre-optimized prompt templates reduce token usage while maintaining quality.
- Image Generation: Supports both external services (e.g., DALL·E, Stable Diffusion) and built-in generation.
- Doubao TTS: High-quality Mandarin TTS using Doubao (豆包) API for natural narration.
- FFmpeg Composition: Automated video assembly with transitions, subtitles, and timing.
Usage
- Prepare a storyboard or script.
- Use the skill's prompt pack to generate images.
- Generate audio narration via Doubao TTS.
- Compose the final video with FFmpeg.
Example
# Example command to generate a Lianhuanhua video
skill lianhuanhua --script "story.txt" --output "video.mp4"
Requirements
- Python 3.8+
- FFmpeg installed
- Doubao API key (optional for external TTS)
- Image generation service (optional for external generation)
Notes
- Token-saving prompt packs are optimized for Codex and similar AI assistants.
- All generated content should comply with applicable terms of service.