Air-Gapped/Offline Installation
Install and run openclaw.ai in completely isolated environments without internet access using local LLMs.
Security-First
Ideal for secure environments, classified networks, or locations without internet connectivity.
Prepare Packages (Internet Machine)
On a machine WITH internet access, download all required packages:
# Create staging directory and copy the official setup scripts
mkdir -p ~/openclaw-offline
cd ~/openclaw-offline
git clone --depth 1 https://github.com/openclaw/openclaw.git
# Download Ollama (for local LLM)
curl -L https://ollama.com/download/ollama-linux-amd64 -o ollama
# Download LLM model (example: Llama 2 7B Q4)
# Get model file from Hugging Face or other source
# Example: llama-2-7b-chat.Q4_K_M.gguf (~4GB)
# Save the official OpenClaw image
docker pull ghcr.io/openclaw/openclaw:latest
docker save ghcr.io/openclaw/openclaw:latest -o openclaw-image.tar
Place a compatible GGUF model in the staging directory and verify checksums for every transferred artifact.
Transfer to Air-Gapped Machine
Transfer files via USB drive, secure media, or internal network:
# Create archive
cd ~/openclaw-offline
tar -czvf openclaw-offline-bundle.tar.gz .
# Transfer to USB drive
cp openclaw-offline-bundle.tar.gz /media/usb-drive/
Security Review
Follow your organization's data transfer procedures for air-gapped systems.
Prepare the Local Model
Extract the bundle, install Ollama, and import the model before starting OpenClaw:
# Extract bundle
tar -xzvf openclaw-offline-bundle.tar.gz
cd openclaw-offline
# Install and start Ollama on an address Docker can reach
chmod +x ollama
sudo mv ollama /usr/local/bin/
OLLAMA_HOST=0.0.0.0:11434 ollama serve > ollama.log 2>&1 &
# Import the transferred GGUF model
cat > Modelfile << EOF
FROM ./llama-2-7b-chat.Q4_K_M.gguf
EOF
ollama create llama2-local -f Modelfile
ollama list
Run OpenClaw Offline
Load the transferred OpenClaw image and run the official setup script in offline mode:
docker load -i openclaw-image.tar
cd openclaw
export OPENCLAW_IMAGE="ghcr.io/openclaw/openclaw:latest"
./scripts/docker/setup.sh --offline
During onboarding, choose Ollama, use http://host.docker.internal:11434 as the base URL,
and select llama2-local. Do not append /v1.
Completely Offline!
openclaw.ai is now running with a local LLM, requiring no internet connectivity.
Was this helpful?