🎬 VideoTranscriberPro

VideoTranscriberPro is a lightweight web application that allows users to upload video files and receive automatic transcriptions using OpenAI's Whisper model. With a sleek frontend, Whisper-powered backend, and support for multiple file formats, it’s the perfect tool for fast, high-quality video transcription.

---

🚀 Features

• 🎥 Upload multiple videos at once (drag & drop interface)
• 🧠 Automatic transcription with OpenAI Whisper
• 📝 Output with and without timestamps
• 🌐 Downloadable SRT subtitle support
• 📁 Local file handling for privacy and performance
• 🎨 Clean, responsive frontend UI (HTML/CSS/JS)
• 🧪 Optional English translation support (when configured)

---

🗂 Project Structure

VideoTranscriberPro/

├── app.py # Flask backend

├── requirements.txt # Python dependencies

├── templates/

│ └── index.html # Frontend UI

├── static/

│ ├── style.css # Styles

│ └── script.js # Frontend logic

├── uploads/ # Temporary video uploads

└── output/ # Transcription results

yaml Copy Edit

---

⚙️ Installation & Setup

1. Prerequisites

• Python 3.8+
• FFmpeg (must be installed and added to PATH)
  • Windows: Download FFmpeg
  • macOS: brew install ffmpeg
  • Ubuntu: sudo apt-get install ffmpeg

2. Installation Steps

# Clone the repository
git clone https://github.com/mobius29er/VideoTranscriberPro.git
cd VideoTranscriberPro

# (Recommended) Create a virtual environment
python -m venv venv
# Activate (Windows)
venv\Scripts\activate
# Activate (macOS/Linux)
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the Flask app
python app.py
Open your browser to http://localhost:5000

🧪 Usage

• Open the web app
• Drag & drop or select one or more videos
• Click Start Transcription
• Wait for processing (progress bar shows status)
• Download the result (with or without timestamps, SRT) (File will autogenerate also in the output folder for with, without, .srt. language, etc.)

📄 Output Files For each uploaded video, you’ll get:

• ✅ video_transcript.txt – raw transcript (no timestamps)
• ✅ video_with_timestamps.txt – readable transcript with [HH:MM:SS - HH:MM:SS] markers
• ✅ video.srt – subtitle file
• ✅ (Optional) English translation .txt and .srt if translation is enabled

🧠 Whisper Model Options Whisper supports several models:

Model Size Speed Accuracy tiny ~39 MB Very Fast Low base ~74 MB Fast Moderate small ~244 MB Medium Good medium ~769 MB Slower Very Good large ~1550 MB Slowest Best

To change the model, edit app.py:

model = whisper.load_model("medium")  # or "small", "large", etc.
🛠 Configuration
UPLOAD_FOLDER and OUTPUT_FOLDER: Can be changed in app.py

MAX_CONTENT_LENGTH: Controls max upload size (default: 500MB)

ALLOWED_EXTENSIONS: Adjust to accept more/less video types

Translations: Can be enabled if you add a translation module or flag

🧾 Example Output (with timestamps)

[00:00:00 - 00:00:05] Welcome to our demo on AI-powered transcription.
[00:00:05 - 00:00:10] In this video, we’ll explore how Whisper works.

💻 Tech Stack

• Flask – lightweight Python web framework
• Whisper – OpenAI speech recognition model
• JavaScript/CSS – frontend interactivity and styling
• FFmpeg – video to audio conversion tool

📦 Requirements

--extra-index-url https://download.pytorch.org/whl/nightly/cu128

Flask==3.0.0
openai-whisper
torch==2.9.0.dev20250716+cu128
torchvision==0.24.0.dev20250717+cu128
torchaudio==2.8.0.dev20250717+cu128
werkzeug==3.0.1
numpy<2

pip install -r requirements.txt

💡 Acknowledgements

Special thanks to PyVideoTrans for their excellent guide on enabling PyTorch Blackwell (RTX 5xxx series) support via CUDA 12.8 Nightly.

• CUDA 12.8 + Nightly Torch for RTX 5xxx (Blackwell) support
• Based on guidance from: https://pyvideotrans.com/en/blog/5090s

📌 Notes

• First-time run will auto-download the Whisper model
• FFmpeg must be correctly installed for audio extraction to work
• Your machine must support the chosen Whisper model (larger models require more memory)

📣 Future Features (Planned)

• ✅ Language detection & translation toggle
• ✅ SRT subtitle preview in browser
• ✅ GPU support via PyTorch CUDA if available
• ⏳ User authentication (multi-user support)
• ⏳ Cloud deployment template (Render, Vercel, Heroku)

🤝 Contributing

Contributions are welcome!

Please open an issue or submit a PR with any enhancements, fixes, or ideas.

📜 License

This project is licensed under the MIT License.

👤 Author

Jeremy Foxx

🎉 Love this tool? Help support ongoing development: Sponsor me on GitHub