GitHub Docs

Deploy VibeCody on Linux

Supports Ubuntu, Debian, Fedora, Arch, and most distributions. Optional CUDA/ROCm GPU acceleration.

Setup time: 2 minutes Cost: Free GPU: CUDA (NVIDIA) / ROCm (AMD)

Quick Start

curl -fsSL https://raw.githubusercontent.com/TuringWorks/vibecody/main/install.sh | sh
vibecli

Or the full setup with always-on service:

cd vibecody/deploy/linux
./setup.sh --always-on

Step-by-Step

1. Install Dependencies

# Ubuntu/Debian
sudo apt install curl pkg-config libssl-dev

# Fedora
sudo dnf install curl openssl-devel

# Arch
sudo pacman -S curl openssl

2. Install VibeCLI

curl -fsSL https://raw.githubusercontent.com/TuringWorks/vibecody/main/install.sh | sh

3. Install Ollama (Optional)

curl -fsSL https://ollama.com/install.sh | sh
ollama pull codellama:7b

4. Enable Always-On (Optional)

# Using our setup script:
./setup.sh --always-on

# Or manually:
mkdir -p ~/.config/systemd/user
cp deploy/linux/vibecody.service ~/.config/systemd/user/
systemctl --user daemon-reload
systemctl --user enable --now vibecody.service

Check status: systemctl --user status vibecody

GPU Setup

NVIDIA (CUDA)

# Install NVIDIA drivers and container toolkit
sudo apt install nvidia-driver-535 nvidia-container-toolkit
ollama pull codellama:7b  # Ollama auto-detects CUDA

AMD (ROCm)

sudo apt install rocm-dev
# Ollama supports ROCm on compatible GPUs

Uninstall

./uninstall.sh

Troubleshooting

Problem Solution
TLS errors building from source Install pkg-config and libssl-dev
Service fails to start Check: journalctl --user -u vibecody -f
Ollama not using GPU Verify: nvidia-smi or check /dev/kfd for AMD
Binary not found Add ~/.local/bin to PATH

What’s Next