Video Generator | 视频生成器

SKILL.md

---
name: video-generator
description: Automated text-to-video pipeline with multi-provider TTS/ASR support - OpenAI, Azure, Aliyun, Tencent | 多厂商 TTS/ASR 支持的自动化文本转视频系统
tags: [video-generation, remotion, openai, azure, aliyun, tencent, tts, whisper, automation, ai-video, short-video, text-to-video, multi-provider]
repository: https://github.com/ZhenRobotics/openclaw-video-generator
homepage: https://github.com/ZhenRobotics/openclaw-video-generator#readme
requires:
  api_keys:
    - name: OPENAI_API_KEY
      description: OpenAI API key for TTS and Whisper services (default provider) | OpenAI API 密钥（默认提供商）
      url: https://platform.openai.com/api-keys
      optional: false
      provider_group: openai
    - name: AZURE_SPEECH_KEY
      description: Azure Speech Service key (alternative to OpenAI) | Azure 语音服务密钥（OpenAI 的替代选项）
      url: https://portal.azure.com/
      optional: true
      provider_group: azure
    - name: AZURE_SPEECH_REGION
      description: Azure region (required with AZURE_SPEECH_KEY) | Azure 区域（配合 AZURE_SPEECH_KEY 使用）
      optional: true
      provider_group: azure
    - name: ALIYUN_ACCESS_KEY_ID
      description: Aliyun AccessKey ID (alternative to OpenAI) | 阿里云 AccessKey ID（OpenAI 的替代选项）
      url: https://ram.console.aliyun.com/manage/ak
      optional: true
      provider_group: aliyun
    - name: ALIYUN_ACCESS_KEY_SECRET
      description: Aliyun AccessKey Secret (required with ALIYUN_ACCESS_KEY_ID) | 阿里云 AccessKey Secret
      optional: true
      provider_group: aliyun
    - name: ALIYUN_APP_KEY
      description: Aliyun NLS App Key (required with ALIYUN_ACCESS_KEY_ID) | 阿里云 NLS 应用密钥
      url: https://ram.console.aliyun.com/manage/ak
      optional: true
      provider_group: aliyun
    - name: TENCENT_SECRET_ID
      description: Tencent Cloud Secret ID (alternative to OpenAI) | 腾讯云 Secret ID（OpenAI 的替代选项）
      url: https://console.cloud.tencent.com/cam/capi
      optional: true
      provider_group: tencent
    - name: TENCENT_SECRET_KEY
      description: Tencent Cloud Secret Key (required with TENCENT_SECRET_ID) | 腾讯云 Secret Key
      optional: true
      provider_group: tencent
    - name: TENCENT_APP_ID
      description: Tencent Cloud App ID (required with TENCENT_SECRET_ID) | 腾讯云应用 ID
      optional: true
      provider_group: tencent
  tools:
    - node>=18
    - npm
    - ffmpeg
    - python3
    - jq
  packages:
    - name: openclaw-video-generator
      source: npm
      version: "1.6.2"
      registry_url: https://www.npmjs.com/package/openclaw-video-generator
      verified_repo: https://github.com/ZhenRobotics/openclaw-video-generator
      verified_commit: 6279034
      npm_latest: "1.6.2"
install:
  method: npm
  package: openclaw-video-generator@1.6.2
  commands:
    - npm install -g openclaw-video-generator@1.6.2
  verify:
    - openclaw-video-generator --version
  registry_verification: |
    npm info openclaw-video-generator version
    # Expected output: 1.6.2
  post_install: |
    Configure ONE provider (see api_keys section above):
    OpenAI (default): export OPENAI_API_KEY="sk-..."
    Or Azure/Aliyun/Tencent (see api_keys for details)
  notes: |
    Install spec note: npm package.json does not include environment variable declarations
    (this is a spec limitation, not a security issue). All configuration is documented above
    in the api_keys section. Verification: npm info openclaw-video-generator
---

# 🎬 Video Generator Skill

Automated text-to-video generation system that transforms text scripts into professional short videos with AI-powered voiceover, precise timing, and cyber-wireframe visuals.

**Cost**: ~$0.003 per 15-second video | **License**: MIT | **Package**: [openclaw-video-generator](https://www.npmjs.com/package/openclaw-video-generator)

---

## 📦 Package Information

| Property | Value |
|----------|-------|
| npm Package | `openclaw-video-generator` |
| Version | 1.6.2 |
| Repository | [github.com/ZhenRobotics/openclaw-video-generator](https://github.com/ZhenRobotics/openclaw-video-generator) |
| Commit Hash | `6279034` |
| License | MIT |

**Verification**:
```bash
npm info openclaw-video-generator version repository.url
# Expected: 1.6.2 and https://github.com/ZhenRobotics/openclaw-video-generator
```

---

## 🔐 Provider Setup (Choose ONE)

This tool supports **4 alternative TTS/ASR providers**. You only need **ONE** configured:

### Option 1: OpenAI (Recommended)
```bash
export OPENAI_API_KEY="sk-..."
```
- Pros: Best quality, simple setup
- Cost: ~$0.003 per 15s video

### Option 2: Azure
```bash
export AZURE_SPEECH_KEY="..."
export AZURE_SPEECH_REGION="eastasia"
```
- Pros: Enterprise reliability
- Cost: Similar to OpenAI

### Option 3: Aliyun (阿里云)
```bash
export ALIYUN_ACCESS_KEY_ID="..."
export ALIYUN_ACCESS_KEY_SECRET="..."
export ALIYUN_APP_KEY="..."
```
- Pros: China connectivity, Chinese voices
- Cost: ~¥0.02 per 15s video

### Option 4: Tencent (腾讯云)
```bash
export TENCENT_SECRET_ID="..."
export TENCENT_SECRET_KEY="..."
export TENCENT_APP_ID="..."
```
- Pros: China connectivity
- Cost: ~¥0.02 per 15s video

**Why multiple providers?** Fallback support for network restrictions, regional preferences, and cost optimization.

---

## 🚀 Quick Start

### Prerequisites
```bash
node --version  # Need >= 18
npm --version
ffmpeg -version
```

### Installation

**Option 1: npm Global Install**
```bash
npm install -g openclaw-video-generator@1.6.2
export OPENAI_API_KEY="sk-..."  # Or add to ~/.bashrc
openclaw-video-generator --version
```

**Option 2: From Source**
```bash
git clone https://github.com/ZhenRobotics/openclaw-video-generator.git
cd openclaw-video-generator
npm install

# Configure provider
cp .env.example .env
nano .env  # Add your API key
chmod 600 .env
```

### First Video
```bash
cd ~/openclaw-video-generator
cat > test.txt << 'EOF'
AI makes development easier
Saving time and boosting efficiency
EOF

./scripts/script-to-video.sh test.txt --voice nova --speed 1.15
# Output: out/test.mp4
```

---

## 💻 Agent Usage

### When to Use
Auto-trigger when user mentions: `video`, `generate video`, `create video`, `生成视频`

### Standard Command
```bash
cd ~/openclaw-video-generator && \
./scripts/script-to-video.sh <script-file> \
  --voice nova \
  --speed 1.15
```

### With Background Video
```bash
cd ~/openclaw-video-generator && \
./scripts/script-to-video.sh <script-file> \
  --voice nova \
  --bg-video "backgrounds/tech.mp4" \
  --bg-opacity 0.6
```

### Example Flow

User: "Generate video: AI makes development easier"

Agent:
```bash
# 1. Check project
ls ~/openclaw-video-generator || echo "Not installed"

# 2. Create script
cat > ~/openclaw-video-generator/scripts/user-script.txt << 'EOF'
AI makes development easier
EOF

# 3. Generate
cd ~/openclaw-video-generator && \
./scripts/script-to-video.sh scripts/user-script.txt

# 4. Show result
echo "Video: ~/openclaw-video-generator/out/user-script.mp4"
```

### Guidelines

**Do**:
- Verify project exists before running
- Check .env configuration
- Show output file location

**Don't**:
- Clone without user confirmation
- Hardcode API keys in commands
- Create new Remotion projects

---

## 🎯 Core Features

- **Multi-Provider TTS**: OpenAI, Azure, Aliyun, Tencent with auto-fallback
- **Timestamp Extraction**: Precise speech-to-text segmentation
- **Scene Detection**: 6 intelligent scene types with auto-styling
- **Video Rendering**: Remotion with cyber-wireframe aesthetics
- **Background Videos**: Custom backgrounds with opacity control
- **Local Processing**: Video rendering happens on your machine

---

## ⚙️ Configuration

### TTS Voices

**OpenAI**:
- `nova` (recommended), `alloy`, `echo`, `shimmer`

**Azure**:
- `zh-CN-XiaoxiaoNeural`, `zh-CN-YunxiNeural`

### Speech Speed
Range: 0.25 - 4.0 | Recommended: 1.15

### Background Video
- `--bg-video <path>` - Video file
- `--bg-opacity <0-1>` - Transparency
- `--bg-overlay <rgba>` - Text overlay

**Recommended**:
| Use Case | Opacity | Overlay |
|----------|---------|---------|
| Text-focused | 0.3-0.4 | `rgba(10,10,15,0.6)` |
| Balanced | 0.5-0.6 | `rgba(10,10,15,0.4)` |
| Visual-focused | 0.7-1.0 | `rgba(10,10,15,0.25)` |

---

## 📊 Video Specs

- Resolution: 1080 x 1920 (vertical)
- Frame Rate: 30 fps
- Format: MP4 (H.264 + AAC)
- Style: Cyber-wireframe with neon colors
- Duration: Auto-calculated

---

## 🎨 Scene Types

| Type | Effect | Trigger |
|------|--------|---------|
| title | Glitch + scale | First segment |
| emphasis | Pop-up zoom | Numbers/percentages |
| pain | Shake + warning | Problems mentioned |
| content | Fade-in | Regular text |
| circle | Rotating ring | Listed points |
| end | Slide-up | Last segment |

---

## 💰 Cost

Per 15-second video: **~$0.003** (< 1 cent)
- TTS: ~$0.001
- Whisper: ~$0.0015
- Rendering: Free (local)

---

## 🔧 Troubleshooting

### Project Not Found
```bash
ls ~/openclaw-video-generator || \
git clone https://github.com/ZhenRobotics/openclaw-video-generator.git ~/openclaw-video-generator && \
cd ~/openclaw-video-generator && npm install
```

### API Key Error
```bash
# Verify .env
cat ~/openclaw-video-generator/.env

# Create if missing
cd ~/openclaw-video-generator
echo 'OPENAI_API_KEY="sk-..."' > .env
chmod 600 .env
```

### Provider Test
```bash
cd ~/openclaw-video-generator && ./scripts/test-providers.sh
```

---

## 🔒 Privacy

**Local Processing**:
- Video rendering
- Scene orchestration
- File management

**Cloud Processing** (via configured provider):
- Text-to-Speech (text sent to API)
- Speech recognition (audio sent to API)

API keys are stored in `.env` file (600 permissions, never committed to git).

---

## 📚 Documentation

- **npm**: https://www.npmjs.com/package/openclaw-video-generator
- **GitHub**: https://github.com/ZhenRobotics/openclaw-video-generator
- **Issues**: https://github.com/ZhenRobotics/openclaw-video-generator/issues

---

## 📊 Tech Stack

Remotion · OpenAI · Azure · Aliyun · Tencent · TypeScript · Node.js · FFmpeg

---

## 🆕 Version History

**v1.6.2** (2026-03-25) - Current
- Chinese TTS integration (Aliyun)
- Dual subtitle styles
- Medical content examples

**v1.6.0** (2026-03-18)
- Premium styles system
- Poster generator
- Design tokens

**v1.2.0** (2026-03-07)
- Background video support
- Multi-provider architecture
- Auto-fallback

**v1.0.0** (2026-03-03)
- Initial release

---

**License**: MIT | **Author**: @ZhenStaff | **Support**: [GitHub Issues](https://github.com/ZhenRobotics/openclaw-video-generator/issues)