An AI-powered app for summarizing meeting videos working fully locally.
- macOS (Apple Silicon, at least 32GB RAM)
- Approximately 20GB of free disk space (for model download)
-
Clone this repository.
-
Make the installation script executable:
chmod +x install.sh
-
Run the installation script:
./install.sh
This script will:
- Install Homebrew (if not already installed)
- Install Python 3.10, Go, ffmpeg, and wget
- Create a Python virtual environment
- Install all required Python packages
- Download the Qwen2.5 model (~20GB)
- Build the Go application
-
Make the run script executable (only needed once):
chmod +x run.sh
-
Start the application:
./run.sh
If you encounter any issues:
-
Make sure all prerequisites are installed:
brew install [email protected] go ffmpeg wget
-
If the model download fails, you can manually download it:
wget -O models/Qwen2.5-32B-Instruct-Q4_K_M.gguf \ "https://huggingface.co/bartowski/Qwen2.5-32B-Instruct-GGUF/resolve/main/Qwen2.5-32B-Instruct-Q4_K_M.gguf?download=true"
-
Try rebuilding the application:
source venv/bin/activate go build -o summarizer_server cmd/server/main.go
-
Check if the Python virtual environment is activated:
source venv/bin/activate
MIT