kids-word-story-video 是做什么的?
A Codex skill for creating children's English word story videos with stories, illustrations, narration, word overlays, sound effects, and Remotion.
This skill is designed for Chinese-speaking users of Codex (or compatible AI coding agents) to automatically produce engaging children's English word story videos. It integrates multiple creative components: generating a short story around a target English word, creating matching illustrations, producing AI narration (TTS), adding word overlays for vocabulary emphasis, incorporating sound effects, and rendering the final video using Remotion. The workflow is fully automated, making it ideal for educators, content creators, or parents who want to quickly produce educational video content.
How It Works
- Story Generation: You provide a target English word. The skill prompts an AI language model to generate a short, age-appropriate story that highlights the word.
- Illustration Creation: Based on the story, the skill calls an image generation model (e.g., DALL-E, Stable Diffusion) to create a series of illustrations.
- Narration & Sound Effects: Text-to-speech (TTS) is used to narrate the story in clear English. Optional sound effects can be added to enhance engagement.
- Word Overlays & Composition: The skill adds animated word overlays to emphasize the target vocabulary. All assets are combined into a Remotion project.
- Final Render: Remotion renders the final MP4 video, ready for sharing or publishing.
Supported Platforms
- Codex (primary target)
- Anthropic Claude Code (compatible)
- Any AI coding agent that supports executing shell commands and API calls
Prerequisites
- Node.js 18+
- Remotion installed
- Access to an AI model API (for story & image generation)
- TTS service (e.g., Google Cloud TTS, ElevenLabs)
Usage
-
Set up your environment:
npm install # or yarn install -
Configure API keys (set via environment variables):
OPENAI_API_KEY=your_key_here TTS_API_KEY=your_tts_key_here -
Run the skill with a target word:
npx remotion render src/index.ts kids_word_story_video --props='{"word": "apple"}'
Output
The skill generates an MP4 video file containing:
- A narrated story with the target word prominently featured
- Matching illustrations for each scene
- Word overlays (e.g., "Let's learn the word 'apple'!")
- Background music or sound effects (optional)
- End credits with the word and its meaning
Customization
You can modify the config.ts file to adjust:
- Story length (number of scenes)
- Illustration style (e.g., "watercolor", "cartoon")
- Voice type (e.g., "female", "child-friendly")
- Video resolution and framerate
Example
For the word "garden":
- The story might follow a child planting seeds and watching flowers grow.
- Illustrations show a seed, sprout, flower, and butterfly.
- Narration emphasizes the word "garden" in context.
- Word overlay appears when the word is spoken.
Notes
- Ensure you have sufficient API credits for image generation and TTS.
- The first render may take a few minutes to cache assets.
- This skill is in alpha; feedback and contributions are welcome!