This guide walks through installing and setting up LLM Harbor for local LLM stack management.
For Docker installation, see Docker.
Harbor can be installed via multiple methods:
curl -fsSL https://raw.githubusercontent.com/av/harbor/main/install.sh | bash
pip install llm-harbor
# Or using pipx for isolated installation
pipx install llm-harbor
npm install -g @avcodes/harbor
git clone https://github.com/av/harbor.git ~/.harbor
cd ~/.harbor
./harbor.sh ln
harbor --version
harbor doctor
The harbor doctor command checks system compatibility and should show:
Start the default stack (Open WebUI + Ollama):
harbor up
Open Harbor in your browser or use the QR code for mobile access:
harbor open
# Or print QR code
harbor qr
# Add Web RAG with search
harbor up searxng
# Add voice chat (STT/TTS)
harbor up speaches
# Multiple backends
harbor up ollama llamacpp vllm
# Full workflow platform
harbor up dify
# Image generation with FLUX
harbor up comfyui
Prefer automation? See LLM Harbor Ansible Setup for an example playbook.
Prefer containers? See LLM Harbor Docker Setup for Docker Compose usage.
Any questions?
Feel free to contact us. Find all contact information on our contact page.