Tutorial Video Generator

AI-powered tutorial video generator - Automatically create professional tutorial videos from documentation and code.

✨ Features

🤖 Fully Automated - AI reads docs, plans tutorial, records operations, generates narration
🎬 Professional Quality - 1080p videos with synchronized voiceover, highlights, and transitions
🆓 Completely Free - Uses local models (CSGLite) + free TTS (Edge TTS) - NO API keys required
🌍 Multilingual - Chinese and English narration support
🎯 Multiple Scenarios - Quickstart, Feature Tour, Developer Integration guides
🔒 Privacy-First - All processing can happen locally

🚀 Quick Start (Fully Local - No API Keys)

Prerequisites

CSGLite - Running locally
```
csghub-lite serve
```

Basic Dependencies

# Already included in Claude Code:
# - FFmpeg ✓
# - Playwright ✓

Installation

# Clone the repository
git clone https://github.com/frankfika/tutorial-video-gen.git
cd tutorial-video-gen

# Run setup (installs Edge TTS in virtual environment)
./install.sh

Generate Your First Video

# Activate the environment
source venv/bin/activate

# Generate a quickstart tutorial
/tutorial-video-gen quickstart --repo ~/your-project --lang zh

That's it! Your video will be ready in ~6-10 minutes.

📹 What Gets Generated

Each video includes:

✅ Professional title card
✅ Step-by-step demonstrations (6-8 steps)
✅ Natural narration (Chinese/English)
✅ Visual highlights on clicks
✅ Smooth transitions
✅ Resource links and outro
✅ 1920x1080 Full HD, ~40-50MB

Example Output Structure

0:00 - 0:05   Title: "Project Name - Quick Start"
0:05 - 0:20   Installation steps
0:20 - 0:50   First run demonstration
0:50 - 1:10   Basic interaction
1:10 - 1:35   UI overview
1:35 - 2:10   Common workflows
2:10 - 2:20   Next steps & resources

🎯 Supported Scenarios

1. Quickstart Tutorial (2-3 minutes)

Perfect for new users - installation, configuration, first example

/tutorial-video-gen quickstart --repo ~/my-app

2. Feature Tour (5-8 minutes)

Comprehensive overview of all main features

/tutorial-video-gen feature-tour --url http://localhost:3000

3. Developer Integration (5-10 minutes)

API documentation and code examples for developers

/tutorial-video-gen developer --repo ~/my-sdk --focus "REST API"

⚙️ Configuration Modes

Mode 1: Fully Local (Recommended)

NO API keys needed!

# .env configuration
MODE=local
CSGLITE_URL=http://localhost:11435
CSGLITE_MODEL=Qwen/Qwen2.5-7B-Instruct
TTS_PROVIDER=edge-tts

Pros:

✅ Completely free
✅ Full privacy (no data leaves your machine)
✅ Unlimited usage
✅ No API quotas

Requirements:

CSGLite running locally with 7B+ model
Edge TTS installed (pip install edge-tts)

Mode 2: Hybrid

Mix local and cloud for best balance

# Option A: Local LLM + Cloud TTS
MODE=hybrid
CSGLITE_URL=http://localhost:11435
OPENAI_API_KEY=sk-...
TTS_PROVIDER=openai

# Option B: Cloud LLM + Local TTS
DEEPSEEK_API_KEY=sk-...
TTS_PROVIDER=edge-tts

Mode 3: Cloud

Highest quality, requires API keys

MODE=cloud
OPENAI_API_KEY=sk-...
TTS_PROVIDER=openai

Cost: ~$0.15-$0.35 per video

See CONFIG_GUIDE.md for detailed configuration options.

🛠 How It Works

5-Phase Workflow

1. Analyze (30s-1min)
   ├─ AI reads README and code
   ├─ Extracts features and workflows
   └─ Understands project structure

2. Plan (1-2min)
   ├─ Generates detailed tutorial script
   ├─ Plans 6-8 demonstration steps
   └─ Writes narration for each step

3. Record (2-3min)
   ├─ Playwright executes browser operations
   ├─ asciinema records terminal commands
   └─ Adds visual highlights

4. Narrate (1-2min)
   ├─ AI polishes narration text
   └─ TTS converts to speech

5. Compose (1-2min)
   ├─ FFmpeg combines all clips
   ├─ Overlays audio and effects
   └─ Exports final MP4

Total: ~6-10 minutes

📚 Documentation

Quick Start Guide - Get started in 5 minutes
Configuration Guide - Local vs Cloud setup
Implementation Summary - Technical details

🧪 Testing

# Test all dependencies
python3 test.py

# Test local LLM connection
python3 utils/llm_provider.py

# Test TTS providers
python3 utils/local_tts.py

📖 Examples

Generate CSGLite Tutorial

/tutorial-video-gen quickstart \
  --repo ~/csglite \
  --lang zh \
  --output csglite-quickstart.mp4

Generate Web App Tour

/tutorial-video-gen feature-tour \
  --url http://localhost:3000 \
  --duration 6 \
  --lang en

Generate API Integration Guide

/tutorial-video-gen developer \
  --repo ~/my-sdk \
  --api-docs ~/my-sdk/docs/api \
  --focus "Authentication"

🎨 Customization

Change Narration Style

Edit templates/narration-styles.json:

casual - Friendly, conversational
formal - Professional, structured
technical - Expert-level terminology
beginner - Encouraging, slow-paced

Change Video Quality

# Higher quality (larger file)
VIDEO_QUALITY=18 /tutorial-video-gen quickstart

# Lower quality (smaller file)
VIDEO_QUALITY=28 /tutorial-video-gen quickstart

Change Voice

# For Chinese
TTS_VOICE_ZH=zh-CN-YunxiNeural  # Male
TTS_VOICE_ZH=zh-CN-XiaoxiaoNeural  # Female (default)

# For English
TTS_VOICE_EN=en-US-GuyNeural  # Male
TTS_VOICE_EN=en-US-JennyNeural  # Female (default)

🔧 Troubleshooting

"CSGLite not running"

csghub-lite serve

"Edge TTS not found"

source venv/bin/activate
pip install edge-tts

"FFmpeg not found"

# macOS
brew install ffmpeg

# Linux
sudo apt install ffmpeg

"Playwright browser not installed"

python3 -m playwright install chromium

💡 Pro Tips

Use 7B+ models for best results with local LLM
Edge TTS is free and nearly matches Azure Speech quality
Batch generate tutorials without API limits using local mode
Start with quickstart scenario to verify setup
Keep README updated for better AI analysis

🌟 Why This Project?

📹 Save Time - Generate tutorials automatically instead of manual recording
🔄 Stay Updated - Regenerate tutorials when code changes
🌍 Reach More Users - Easy to create multilingual versions
💰 Zero Cost - Completely free when using local mode
🔒 Privacy - Your code never leaves your machine (local mode)

📊 Provider Comparison

Provider	Cost	Quality	Privacy	Installation
CSGLite + Edge TTS	Free	Very Good	✅ Local	Easy
DeepSeek + Edge TTS	$0.14/video	Very Good	⚠️ Cloud LLM	Easy
OpenAI (both)	$0.35/video	Best	⚠️ Cloud	Easy

🤝 Contributing

Contributions welcome! Areas for improvement:

📄 License

MIT License - see LICENSE

🙏 Acknowledgments

Built with:

Claude Code - AI orchestration and workflow
Playwright - Browser automation
FFmpeg - Video processing
Edge TTS - Free text-to-speech
CSGLite - Local LLM inference

🔗 Links

Made with ❤️ for the open source community

Generate your first tutorial video in under 10 minutes - completely free!

Name		Name	Last commit message	Last commit date
examples		examples
templates		templates
utils		utils
workflows		workflows
.env.template		.env.template
.gitignore		.gitignore
CONFIG_GUIDE.md		CONFIG_GUIDE.md
IMPLEMENTATION_SUMMARY.md		IMPLEMENTATION_SUMMARY.md
LICENSE		LICENSE
QUICKSTART.md		QUICKSTART.md
README.md		README.md
install.sh		install.sh
skill.md		skill.md
test.py		test.py

Latest commit

History

Folders and files

Repository files navigation