ReelsMaker - Headless Video Generation

ReelsMaker is a Python-based application designed to create captivating faceless videos for social media platforms like TikTok and YouTube. This version focuses on headless functionality with YouTube integration via MCP Composio.

Features

• AI-Powered Content Generation: Automatically generate creative scripts for your video content
• Text-to-Speech: Use TikTok or ElevenLabs voices to add synthetic narration
• Auto Subtitles: Generate professional subtitles with customizable fonts and styling
• Background Videos: Automatically fetch relevant stock videos from Pexels
• Headless Operation: Run without UI for automation and batch processing
• YouTube Integration: Upload generated videos directly to YouTube via MCP Composio

Installation

git clone <your-repo-url>
cd reelsmaker

Create a virtual environment and install dependencies:

poetry shell
poetry install

Copy and update the environment file:

cp .env.example .env

Edit .env with your API keys:

OPENAI_API_KEY="your_openai_api_key_here"
PEXELS_API_KEY="your_pexels_api_key_here"
TOGETHER_API_KEY="your_together_api_key_here"
SENTRY_DSN=""
OPENAI_MODEL_NAME="gpt-4-turbo"

Download required spaCy model:

poetry run python -m spacy download en_core_web_sm

Required Assets

Background Video Download

For Subway Surfers background videos, download the gameplay video from:

YouTube Link: https://youtu.be/QPW3XwBoQlw?feature=shared

1. Use any YouTube downloader tool (e.g., yt-dlp, online converters)
2. Download the video in MP4 format
3. Place the downloaded file in your project directory
4. The system will automatically use it for background gameplay footage

Note: Ensure the downloaded video file is named appropriately and is accessible in your project directory.

Usage

Headless Video Generation

Run the headless script to generate videos:

poetry run python headless_runner.py

YouTube Integration Setup

1. Go to MCP Composio Dashboard

   • Visit: https://mcp.composio.dev
   • Sign in to your account

2. Authenticate YouTube

   • Go to the dashboard
   • Search for "youtube"
   • Click on YouTube and authenticate with your Google account

3. Create MCP Server

   • After authentication, create a new MCP Server
   • Copy the terminal command shown after creation

4. Run MCP Server

   • Run the copied terminal command in your terminal
   • This will start the MCP server for YouTube integration

5. Use MCP Client

   • Use your MCP client to connect to the server
   • Upload and post the generated videos directly to YouTube

Configuration

You can customize video generation by modifying the ReelsMakerConfig in headless_runner.py:

• prompt: Content theme for your video
• script_duration: Target video length in seconds
• fontsize: Subtitle font size
• font_name: Font family for subtitles
• stroke_color / text_color: Subtitle styling
• watermark_path_or_text: Custom watermark

API Keys Required

• OpenAI: For script generation and content creation
• Pexels: For stock video downloads
• Together AI: For enhanced AI capabilities
• ElevenLabs (optional): For premium voice synthesis

Output

Generated videos are saved in the cache/ directory with the following structure:

• High-quality MP4 format
• 9:16 aspect ratio (optimized for mobile)
• Professional subtitles with custom styling
• Background music support (optional)

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

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