Skip to main content

​
Prerequisites

Before you begin, ensure you have:
  • Python 3.10+ installed
  • An OpenAI API key (get one here)
  • FFmpeg and ImageMagick installed on your system
For detailed installation instructions for all platforms, see the Installation Guide.

​
Quick Setup

1

Clone the Repository

git clone https://github.com/SamurAIGPT/AI-Youtube-Shorts-Generator.git
cd AI-Youtube-Shorts-Generator
2

Install System Dependencies

sudo apt install -y ffmpeg libavdevice-dev libavfilter-dev \
  libopus-dev libvpx-dev pkg-config libsrtp2-dev imagemagick

# Fix ImageMagick security policy (required for subtitles)
sudo sed -i 's/rights="none" pattern="@\*"/rights="read|write" pattern="@*"/' /etc/ImageMagick-6/policy.xml
3

Create Python Virtual Environment

python3.10 -m venv venv
source venv/bin/activate  # On Windows: .\venv\Scripts\Activate.ps1
4

Install Python Dependencies

pip install -r requirements.txt
If you don’t have an NVIDIA GPU, use requirements-cpu.txt instead, but install CPU PyTorch first:
pip install torch --index-url https://download.pytorch.org/whl/cpu
pip install -r requirements-cpu.txt
5

Configure OpenAI API Key

Create a .env file in the project root:
OPENAI_API=your_openai_api_key_here

​
Generate Your First Short

​
Interactive Mode

Run the tool and enter a YouTube URL when prompted: 
./run.sh
You’ll see: 
Session ID: a1b2c3d4
Enter YouTube video URL or local video file path:
Paste a YouTube URL (e.g., https://youtu.be/dQw4w9WgXcQ) and press Enter.

​
Command-Line Mode

Pass the YouTube URL directly as an argument: 
./run.sh "https://youtu.be/dQw4w9WgXcQ"

​
Using Local Video Files

Process a local video file instead: 
./run.sh "/path/to/your/video.mp4"

​
The Processing Workflow

1

Resolution Selection

You’ll see available video streams:
Available video streams:
  0. Resolution: 1080p, Size: 45.2 MB, Type: Adaptive
  1. Resolution: 720p, Size: 28.1 MB, Type: Adaptive
  2. Resolution: 480p, Size: 15.3 MB, Type: Adaptive

Select resolution number (0-2) or wait 5s for auto-select...
Auto-selecting highest quality in 5 seconds...
  • Enter a number to select immediately
  • Wait 5 seconds to auto-select highest quality
2

Transcription & AI Analysis

The tool will:
  • Extract and transcribe audio (~30s for 5-minute video with GPU)
  • Analyze transcript with GPT-4o-mini to find engaging segments
Analyzing transcription to find best highlight...
3

Review & Approve Selection

The AI will present its selection:
============================================================
SELECTED SEGMENT DETAILS:
Time: 68s - 187s (119s duration)
============================================================

Options:
  [Enter/y] Approve and continue
  [r] Regenerate selection
  [n] Cancel

Auto-approving in 15 seconds if no input...
Your options:
  • Press Enter or y to approve
  • Press r to regenerate (can repeat multiple times)
  • Press n to cancel
  • Wait 15 seconds to auto-approve
4

Video Processing

The tool will:
Step 1/4: Extracting clip from original video...
Step 2/4: Cropping to vertical format (9:16)...
Step 3/4: Adding subtitles to video...
Step 4/4: Adding audio to final video...

​
Expected Output

When processing completes, you’ll see: 
============================================================
✓ SUCCESS: my-awesome-video_a1b2c3d4_short.mp4 is ready!
============================================================

Cleaned up temporary files for session a1b2c3d4
Your vertical short will be saved in the project directory with:
  • 9:16 aspect ratio (perfect for TikTok/Reels/Shorts)
  • Stylized subtitles with Franklin Gothic font
  • Smart cropping (face-centered or motion-tracked)
  • Original audio from selected segment

​
Automation for Batch Processing

Skip interactive prompts for hands-free operation: 
./run.sh --auto-approve "https://youtu.be/VIDEO_ID"
Process multiple videos from a file: 
# Create urls.txt with one URL per line
echo "https://youtu.be/VIDEO1" > urls.txt
echo "https://youtu.be/VIDEO2" >> urls.txt
echo "https://youtu.be/VIDEO3" >> urls.txt

# Process all videos sequentially
xargs -a urls.txt -I{} ./run.sh --auto-approve {}
The --auto-approve flag automatically:
  • Selects highest video quality
  • Approves AI highlight selection after 15s timeout
  • Perfect for overnight batch processing

​
Concurrent Processing

Run multiple instances simultaneously (each gets a unique session ID): 
./run.sh "https://youtu.be/VIDEO1" &
./run.sh "https://youtu.be/VIDEO2" &
./run.sh "/path/to/video3.mp4" &
Each instance operates independently with separate temporary files.

​
Troubleshooting Common Issues

No subtitles appearing?Ensure ImageMagick policy allows file operations:
grep 'pattern="@\*"' /etc/ImageMagick-6/policy.xml
# Should show: rights="read|write"
If not, run:
sudo sed -i 's/rights="none" pattern="@\*"/rights="read|write" pattern="@*"/' /etc/ImageMagick-6/policy.xml
CUDA/GPU errors?The run.sh script automatically configures CUDA library paths. If you still encounter issues:
export LD_LIBRARY_PATH=$(find $(pwd)/venv/lib/python3.10/site-packages/nvidia -name "lib" -type d | paste -sd ":" -)
Or switch to CPU-only mode (see Installation Guide).

​
What’s Next?

Installation Guide

Detailed setup instructions for all platforms including Docker

Configuration

Customize subtitle styling, AI prompts, and video quality settings

Build docs developers (and LLMs) love

Get started for freeTalk to us