The Easiest Free OpenClaw Setup with Ollama’s Cloud Model Support

Step 0: Install an Agent Tool

Install a CLI‑based agent such as Gemini‑CLI, Claude Code, Qwen Coder, or Kimi CLI on the virtual machine. The agent will later help install Ollama and OpenClaw.

Step 1: Install Ollama

Run the official installer: curl -fsSL https://ollama.com/install.sh | sh After installation, start the desired cloud model (e.g., GLM‑4.7:cloud) and log in when prompted.

Step 2: Install OpenClaw

Use npm; first ensure Node.js is available. Install nvm and Node 24:

# Install nvm
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.3/install.sh | bash

# Load nvm
". $HOME/.nvm/nvm.sh"

# Install Node.js 24
nvm install 24

# Verify versions
node -v   # v24.13.0
npm -v    # 11.6.2

Then install OpenClaw via npm.

Step 3: Configure OpenClaw

openclaw onboard --install-daemon

If the onboarding wizard reports errors, run openclaw doctor to fix them, then accept the default prompts.

Step 4: Enable Ollama Model Support

Launch the OpenClaw model with Ollama’s updated launch command: ollama launch openclaw Select the previously started GLM‑4.7:cloud model.

Step 5: Connect an Application

Use the built‑in Telegram channel to interact with OpenClaw (see the official docs at https://docs.openclaw.ai/channels/telegram). The author delegated most integration work to the agent.

Overall, the guide recommends letting the initial agent handle the heavy lifting and suggests running the setup on a well‑provisioned environment such as a Mac Mini or an Ubuntu VM.