YouTubePodcaster

Automated AI-Powered Content Creation Pipeline

This project automates the process of transforming trending Reddit discussions into engaging YouTube videos. It leverages AI tools to generate dialogues, create audio and visuals, and compile everything into a polished video ready for publishing.

---

Overview

The YouTubePodcaster pipeline extracts popular topics from Reddit, generates expert-level dialogues using GPT-3.5 Turbo, synthesizes natural-sounding voices with ElevenLabs, creates visuals using AI image models, and compiles everything into a video with background music. The final video is uploaded to YouTube with automatically generated metadata (title, description, and tags).

---

Features

• Reddit Scraping: Fetches trending topics from Reddit based on comment engagement.
• Dialogue Generation: Uses GPT-3.5 Turbo to create conversational dialogues between two characters (Ahmed and Amna).
• Voice Synthesis: Utilizes ElevenLabs to generate realistic voiceovers for each character.
• Image Generation: Creates visuals using AI models like Stable Diffusion, DALL-E, and others.
• Music Generation: Generates background music using AI models like MusicGen.
• Video Compilation: Combines audio, visuals, and music into a cohesive video.
• YouTube Upload: Automatically uploads the final video to YouTube with generated metadata.

---

How It Works

1. Data Collection:

   • Scrape Reddit for trending topics with high engagement.
   • Extract key points from the discussions.

2. Content Generation:

   • Use GPT-3.5 Turbo to create a dialogue between two experts (Ahmed and Amna).
   • Generate voiceovers for each character using ElevenLabs.

3. Visuals and Audio:

   • Create images using AI models based on topic prompts.
   • Generate background music using AI music models.

4. Video Production:

   • Combine audio, visuals, and music into a single video file.
   • Use video editing tools to ensure smooth transitions and synchronization.

5. Publishing:

   • Automatically generate a title, description, and tags using GPT.
   • Upload the final video to YouTube.

---

Technologies Used

• Natural Language Processing (NLP): GPT-3.5 Turbo for dialogue generation and metadata creation.
• Voice Synthesis: ElevenLabs for realistic voice generation.
• Image Generation: Stable Diffusion, DALL-E, and other Hugging Face models for creating visuals.
• Music Generation: MusicGen and other Hugging Face models for background music.
• Video Editing: MoviePy for combining audio, images, and music.
• Web Scraping: PRAW for Reddit data extraction.
• YouTube API: For uploading videos and managing metadata.

---

Installation

1. Clone the repository:

   git clone https://github.com/ynstf/YouTubePodcaster.git  

2. Install dependencies:

   pip install -r requirements.txt  

3. Configure environment variables:

   • Create a .env file in the root directory and add your API keys:

     # Reddit API credentials  
     CLIENT_ID="your_reddit_client_id"  
     CLIENT_SECRET="your_reddit_client_secret"  
     USER_AGENT="your_user_agent"  
     
     # OpenRouter API settings  
     OPENROUTER_API_KEY="your_openrouter_api_key"  
     OPENROUTER_API_URL="https://openrouter.ai/api/v1/chat/completions"  
     MODEL="gpt-3.5-turbo"  
     
     # ElevenLabs Constants  
     VOICE_ID_MAN="your_voice_id_man"  
     VOICE_ID_WOMAN="your_voice_id_woman"  
     AUDIO_OUTPUT_FILE="podcast/podcast_final.mp3"  
     
     # Hugging Face  
     HUGGING_FACE_TOKEN="your_hugging_face_token"  
     MUSIC_API_URL="https://api-inference.huggingface.co/models/facebook/musicgen-small"  
     IMAGE_API_URLS="https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-dev,https://api-inference.huggingface.co/models/ZB-Tech/Text-to-Image,https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-2"  
     
     # YouTube Upload  
     UPLOAD_TO_YOUTUBE="ON"  
     TOKEN_FILE_YOUTUBE="UPLOAD/token.json"  
     

4. Run the pipeline:

   python main.py  

---

File Structure

YouTubePodcaster/  
├── AUDIO/                  # Audio generation scripts  
├── IMAGE/                  # Image generation scripts  
├── LLM/                    # Dialogue generation scripts  
├── MUSIC/                  # Music generation scripts  
├── REDDIT/                 # Reddit scraping scripts  
├── UPLOAD/                 # YouTube upload scripts  
├── VIDEO/                  # Video compilation scripts  
├── PROMPTS/                # Prompt templates for GPT  
├── podcast/                # Generated audio files and Final video output  
├── images/                 # Generated image files  
├── .env                    # Environment variables  
├── main.py                 # Main script to run the pipeline  
├── README.md               # Project documentation  
└── requirements.txt        # Python dependencies  

---

Future Improvements

• Add support for multiple languages and automated subtitles.
• Improve the quality of generated visuals and music.
• Integrate with additional platforms (e.g., TikTok, Instagram).
• Add a user-friendly interface for non-technical users.

---

Contributing

Contributions are welcome! If you have ideas for improvements or new features, feel free to open an issue or submit a pull request.

---

License

This project is licensed under the MIT License. See the LICENSE file for details.