Skip to content

QUICKSTART.md

Latest commit

 

History

History
196 lines (136 loc) · 4.06 KB

QUICKSTART.md

File metadata and controls

196 lines (136 loc) · 4.06 KB
Raw

Quick Start Guide

Get Voice2Note up and running in 10 minutes!

Prerequisites

Install these first:

  1. Python 3.8+

    python3 --version  # Check if installed

  2. FFmpeg

    # Ubuntu/Debian/WSL
sudo apt update && sudo apt install ffmpeg

# macOS
brew install ffmpeg

Automated Setup (Recommended)

Linux/macOS/WSL

# Run the setup script
bash setup.sh

This will automatically:

  • Create virtual environment
  • Install Python dependencies
  • Clone and build whisper.cpp
  • Download Whisper model
  • Initialize database

Windows (Command Prompt)

# Create virtual environment
python -m venv venv
venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Initialize database
python init_db.py

Then manually install whisper.cpp (see Manual Setup below).

Manual Setup

If the automated setup doesn't work:

1. Install Python Dependencies

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Install whisper.cpp

# Clone repository
git clone https://github.com/ggml-org/whisper.cpp.git
cd whisper.cpp

# Build
cmake -B build
cmake --build build -j --config Release

# Download model (choose one)
bash ./models/download-ggml-model.sh base    # Fast, 142MB
bash ./models/download-ggml-model.sh medium  # Recommended, 1.5GB

cd ..

3. Update Configuration

Edit config.py and update these paths:

WHISPER_PATH = './whisper.cpp/build/bin/whisper-cli'
WHISPER_MODEL_PATH = './whisper.cpp/models/ggml-medium.bin'

4. Initialize Database

python init_db.py

Running the Application

# Activate virtual environment (if not already activated)
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Start the server
python app.py

Open your browser to: http://localhost:5000

First Use

Test with YouTube

  1. Go to http://localhost:5000
  2. Paste this short test video: https://www.youtube.com/watch?v=dQw4w9WgXcQ
  3. Click "Process Video"
  4. Wait 5-10 minutes (depending on your CPU)
  5. View your generated notes!

Test with Local File

  1. Record a short voice memo or use an existing video
  2. Upload it via the web interface
  3. Process and view notes

Troubleshooting

"whisper.cpp not found"

Solution: Update WHISPER_PATH in config.py to point to your whisper.cpp binary:

# Find where whisper-cli is located
find . -name "whisper-cli"

# Update config.py with the correct path

"FFmpeg not found"

Solution: Install FFmpeg:

sudo apt install ffmpeg  # Ubuntu/Debian/WSL
brew install ffmpeg      # macOS

"Ollama API error"

Solution: Check your internet connection. The Ollama API requires internet access.

Processing is slow

Solutions:

  • Use the base model instead of medium (edit config.py)
  • Process shorter videos first
  • Consider GPU acceleration (see README.md)

Next Steps

  • Read the full README.md for detailed documentation
  • Explore the Processing History page
  • Try different video lengths and sources
  • Customize note generation prompts in processors/note_generator.py

Getting Help

System Requirements

Component Minimum Recommended
CPU Dual-core Quad-core or better
RAM 4GB 8GB+
Storage 5GB 10GB+
Internet Required for Ollama API Required for Ollama API

Performance Expectations

Processing a 10-minute video:

  • Audio extraction: 10-30 seconds
  • Transcription (base model): 5-10 minutes
  • Transcription (medium model): 10-20 minutes
  • Note generation: 30-60 seconds
  • Total: ~6-21 minutes

With GPU acceleration:

  • Transcription time can be reduced by 5-10x!

Ready to start? Run python app.py and open http://localhost:5000 🚀