Skip to main content

Getting started

Install​

Install or upgrade the chart from the GitLab OCI registry:

helm upgrade --install openclaw oci://registry.gitlab.com/xrow-public/helm-openclaw/charts/openclaw \
  --version 1.68.0 \
  --create-namespace \
  --namespace openclaw

Configure config.gitlab.repository when the workspace should bootstrap itself from a Git repository profile.

Access the UI​

Read the generated login token from the release namespace:

kubectl get secret -n openclaw openclaw-token -o jsonpath='{.data.token}' | base64 -d

For local access, forward the OpenClaw service to 127.0.0.1:

kubectl port-forward --address 127.0.0.1 -n openclaw deployment/openclaw 18789

Open http://localhost:18789 and sign in with the token. After login, approve connected OpenClaw devices from a shell in the application container:

kubectl exec -n openclaw deployment/openclaw -c openclaw --stdin --tty -- /bin/bash

Connect model providers​

OpenAI example:

openclaw models set openai/gpt-5.5
openclaw models auth login --provider openai --set-default

When prompted for the authorization code or redirect URL, paste the http://localhost:1455/auth/callback request URL from the browser.

GitHub Copilot example:

openclaw models set github-copilot/gpt-5.5
openclaw models auth login --provider github-copilot --method device --set-default

OpenRouter can be configured through Helm secrets:

config:
  secrets:
    - name: OPENROUTER_API_KEY
      value: <key>

Then select a model:

openclaw models set openrouter/nvidia/nemotron-nano-9b-v2:free