# Option 1: Clone directly (recommended)
git clone https://github.com/rmacmorran/whispertranscriber.git whisper-transcriber
cd whisper-transcriber

# Option 2: Create empty directory first, then clone
# mkdir whisper-transcriber
# cd whisper-transcriber
# git clone https://github.com/rmacmorran/whispertranscriber.git .

# Create and activate virtual environment
python -m venv venv
.\venv\Scripts\Activate.ps1

# Verify activation (should show (venv) in prompt)

# Upgrade pip first
python -m pip install --upgrade pip

# Install PyTorch with CUDA support (for NVIDIA GPUs)
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

# Install all other dependencies from requirements.txt
pip install -r requirements.txt

# Test GPU acceleration
python test_whisper.py

# List available audio devices  
python audio_devices.py

# Test audio capture (optional)
python test_audio_buffer.py

# List audio devices to find your device ID
python whisper-transcriber.py --list-devices

# Edit config.yaml to set your audio device
# For VB-Audio Virtual Cable, typically device_index: 31

# Run with default settings (uses config.yaml)
python whisper-transcriber.py

# List available audio devices
python whisper-transcriber.py --list-devices

# Use specific audio device (e.g., VB-Audio Virtual Cable)
python whisper-transcriber.py --device 31

# Use different model size
python whisper-transcriber.py --model small

# Specify language (or auto-detect)
python whisper-transcriber.py --language en

# Test individual components
python test_whisper.py       # Test Whisper + GPU
python test_audio_buffer.py  # Test audio capture
python test_vad_chunker.py   # Test VAD chunking (Ctrl+C to skip live test)

# Use microphone directly
python whisper-transcriber.py --device 1  # Replace with your mic device ID

# Basic real-time transcription (uses config.yaml settings)
python whisper-transcriber.py

# Real-time with VB-Audio Virtual Cable (for capturing app audio)
python whisper-transcriber.py --device 31

# Real-time with microphone
python whisper-transcriber.py --device 1  # Use your mic device ID

# High accuracy real-time (slower but better results)
python whisper-transcriber.py --model small --device 31

# Transcribe a single video file
python whisper-transcriber.py -i "video.mp4" -o "transcript.txt"

# Transcribe with timestamps
python whisper-transcriber.py -i "lecture.mp4" -o "lecture.txt" --timestamps

# Use better model for accuracy
python whisper-transcriber.py -i "interview.wav" -o "interview.txt" -m medium

# Generate SRT subtitle file (auto-detects .srt extension)
python whisper-transcriber.py -i "movie.mp4" -o "movie.srt"

# Force specific language
python whisper-transcriber.py -i "spanish.mp3" -o "spanish.txt" --language es

# Minimal output for batch processing
python whisper-transcriber.py -i "audio.wav" -o "output.txt" -q

# Quiet mode with timestamps
python whisper-transcriber.py -i "video.mp4" -o "video.txt" -q -t

# Transcribe all video/audio files in current directory
.\batch-transcribe.ps1

# Include timestamps in all transcripts
.\batch-transcribe.ps1 -Timestamps

# Generate SRT subtitle files instead of text files
.\batch-transcribe.ps1 -Subtitle

# Use better model for higher accuracy
.\batch-transcribe.ps1 -Model medium -Timestamps

# Custom output directory
.\batch-transcribe.ps1 -OutputDir "transcripts" -Timestamps

# Generate subtitles with large model
.\batch-transcribe.ps1 -Subtitle -Model large -OutputDir "subtitles"

# Force overwrite existing files
.\batch-transcribe.ps1 -Force -Model small -Timestamps

# CPU-only processing (if GPU causes crashes)
.\batch-transcribe.ps1 -CpuOnly -Model medium

# High-quality subtitle generation for video production
.\batch-transcribe.ps1 -Subtitle -Model large -OutputDir "final_subs" -Force

# Fast batch processing for content review
.\batch-transcribe.ps1 -Model base -OutputDir "drafts"

# Mixed content with timestamps for meeting recordings
.\batch-transcribe.ps1 -Timestamps -Model medium -OutputDir "meeting_transcripts"

# Copy batch script to your media folder
copy "C:\path\to\whisper-transcriber\batch-transcribe.ps1" .
.\batch-transcribe.ps1 -Subtitle -Model medium

# Set environment variable for global access
$env:WHISPER_TRANSCRIBER_PATH = "C:\Users\rmacmorran\projects\whisper-transcriber\whisper-transcriber.py"
C:\MyVideos\batch-transcribe.ps1 -Timestamps

# Direct execution with full path
C:\Users\rmacmorran\projects\whisper-transcriber\batch-transcribe.ps1 -Subtitle

# Transcribe all media files in current directory
.\batch-transcribe.ps1

# Include timestamps in transcripts
.\batch-transcribe.ps1 -Timestamps

# Use larger model for better accuracy
.\batch-transcribe.ps1 -Model large -Timestamps

# Custom output directory
.\batch-transcribe.ps1 -OutputDir "my_transcripts" -Force

# Option 1: Copy script to media folder
copy "C:\path\to\whisper-transcriber\batch-transcribe.ps1" .
.\batch-transcribe.ps1 -Timestamps -Model large

# Option 2: Set environment variable (recommended)
$env:WHISPER_TRANSCRIBER_PATH = "C:\path\to\whisper-transcriber\whisper-transcriber.py"
C:\any\folder\batch-transcribe.ps1 -Timestamps

# Option 3: Run with full path
C:\path\to\whisper-transcriber\batch-transcribe.ps1 -Timestamps

.\batch-transcribe.ps1 [OPTIONS]

Options:
  -OutputDir <path>   Output directory (default: current directory)
  -Timestamps         Include timestamps in output
  -Model <size>       Whisper model (tiny/base/small/medium/large)
  -Subtitle           Generate subtitle files (.srt) instead of text files (.txt)
  -Force              Overwrite existing transcripts
  -CpuOnly            Force CPU-only processing (avoids GPU-related crashes)
  -Help               Show help message