A comprehensive microservices-based solution for automatic video captioning, designed for 9:16 format videos (shorts). The system consists of four independent services that can be used together or separately, with a modern web interface for easy interaction.

Auto Captions
├── transcriptions/     # Audio/video transcription service
├── ffmpeg-captions/    # FFmpeg-based subtitle rendering
├── remotion-captions/  # Remotion-based video processing
├── web/                # Web interface for user interaction
├── setup.sh            # Global setup script
└── docker-compose.yml  # Docker orchestration

• Port: 3001
• Purpose: Extract audio from video/audio files and generate transcriptions
• Technology: TypeScript, OpenAI Whisper, Whisper.cpp
• Documentation: transcriptions/README.md

• Port: 3002
• Purpose: Generate captioned videos using FFmpeg with ASS subtitle styling
• Technology: TypeScript, FFmpeg, ASS subtitles
• Documentation: ffmpeg-captions/README.md

• Port: 3003
• Purpose: Create highly customizable captioned videos with Remotion
• Technology: TypeScript, Remotion, React-based styling
• Documentation: remotion-captions/README.md

• Port: 80
• Purpose: User-friendly web interface for the entire caption generation workflow
• Technology: PHP, JavaScript, Tailwind CSS
• Features: File upload, transcription editing, service management, real-time preview
• Documentation: web/README.md

• Node.js 22+
• npm or yarn
• FFmpeg (required for all services)
• PHP 8.4+ (for web interface)
• Docker & Docker Compose (for containerized deployment)

1. Clone the repository:

   git clone <repository-url>
   cd AutoCaptions

2. Start all services:

   docker-compose up -d

3. Access the web interface:

   http://localhost:80
   

1. Clone the repository:

   git clone <repository-url>
   cd AutoCaptions

2. Run global setup:

   chmod +x setup.sh
   ./setup.sh

3. Start services individually:

   # Terminal 1 - Transcriptions
   cd transcriptions && npm run dev
   
   # Terminal 2 - FFmpeg Captions
   cd ffmpeg-captions && npm run dev
   
   # Terminal 3 - Remotion Captions
   cd remotion-captions && npm run dev
   
   # Terminal 4 - Web Interface
   cd web && php -S localhost:80

Once running, the services will be available at:

• Web Interface: http://localhost:80 (Primary user interface)
• Transcriptions API: http://localhost:3001
• FFmpeg Captions API: http://localhost:3002
• Remotion Captions API: http://localhost:3003

1. Open Web Interface: Navigate to http://localhost:80
2. Upload Video: Drag and drop your 9:16 video file
3. Generate Transcription: AI-powered speech-to-text processing
4. Edit Captions: Fine-tune text, timing, and formatting
5. Choose Rendering: Select FFmpeg (fast) or Remotion (advanced)
6. Customize Styling: Fonts, colors, positioning, animations
7. Download Result: Get your captioned video

# 1. Generate transcription
curl -X POST http://localhost:3001/api/transcribe \
  -F "file=@video.mp4" \
  -F "service=whisper-cpp"

# 2. Generate captioned video (FFmpeg)
curl -X POST http://localhost:3002/api/captions/generate \
  -H "Content-Type: application/json" \
  -d '{
    "data": {...},
    "video": "video.mp4"
  }'

# 3. Generate captioned video (Remotion)
curl -X POST http://localhost:3003/render \
  -H "Content-Type: application/json" \
  -d '{
    "video": "video.mp4",
    "transcription": {...},
    "props": {...}
  }'

Each service has its own .env configuration file. After running setup, review and customize:

• transcriptions/.env - Whisper models and API keys
• ffmpeg-captions/.env - FFmpeg paths and output settings
• remotion-captions/.env - Remotion rendering configuration

The web interface provides a settings panel to configure all service URLs:

1. Click the gear icon in the header
2. Update service URLs:
   • Transcriptions: http://localhost:3001 (Docker: http://transcriptions:3001)
   • FFmpeg Captions: http://localhost:3002 (Docker: http://ffmpeg-captions:3002)
   • Remotion Captions: http://localhost:3003 (Docker: http://remotion-captions:3003)
3. Test connections and save

All services include health check endpoints:

# Check individual services
curl http://localhost:3001/health    # Transcriptions
curl http://localhost:3002/health    # FFmpeg Captions
curl http://localhost:3003/health    # Remotion Captions

• ffmpeg-captions service is under the MIT License - see the LICENSE file for details.
• transcriptions service is under the MIT License - see the LICENSE file for details.
• remotion-captions service is under the Remotion License - see the LICENSE file for details.
• web service is under the MIT License - see the LICENSE file for details.

• Include the Remotion service in the web service
• Convert videos to webm in the remotion-captions service rather than h264 to avoid installing Google Chrome (and thus enable the ARM64 build)
• In the transcriptions service, add a fallback to whisper-cpp when openai-whisper is used but the API is down

Built with ❤️