How To Configure Ubuntu + OpenClaw Environment and Connect Local Ollama

1. Pre-Deployment Preparation: Verify Prerequisites & Optimize Environment

Before deploying OpenClaw officially, it is essential to complete the verification and optimization of the basic environment to avoid problems such as missing dependencies and version incompatibility in subsequent installations. This tutorial is based on the Ubuntu system in a VMware virtual machine, with NodeJS, Git environment configuration and Windows Ollama deployment completed in the early stage, which are the core prerequisites for OpenClaw operation and directly determine the success of subsequent deployment.

First, ensure the versions of basic software meet the requirements, as OpenClaw has clear version requirements for NodeJS and Git, and low versions will cause compilation failures. Second, optimize the virtual machine display resolution to facilitate subsequent command line operations and source code file viewing. Finally, plan a file storage path in advance and create a special folder to store OpenClaw source code to avoid file confusion.

1. Step 1: Check NodeJS and Git Versions: Open the terminal, run node -v and git -v to confirm NodeJS ≥ v22.22.1 and Git ≥ 2.51.0; upgrade the software if versions are unqualified.
2. Step 2: Optimize Virtual Machine Resolution: Enter Ubuntu system settings, adjust screen resolution and enlarge font display for clearer operation.
3. Step 3: Create Source Code Directory: Make a new folder named work in the user’s home directory to store the extracted OpenClaw source code for easy path location.
4. Step 4: Prepare Windows Ollama: Enable Ollama network exposure in advance and record the Windows host IP address for cross-end connection.

2. Get Source Code: Download Stable OpenClaw from GitHub

Obtaining OpenClaw source code is the first step of deployment, and choosing the right version is crucial — the latest commit version may have unfixed bugs, so it is recommended to download the official Release stable version to ensure operational stability. This tutorial uses the 2026.3.13 version, which has verified compatibility and stability, making it suitable for beginners.

If GitHub access fails or download gets stuck, you can leave a message to get the downloaded source code package to save time. After downloading, extract the files in time to avoid source code loss due to network interruption during installation.

1. Step 1: Access GitHub via Browser: Launch Ubuntu’s built-in Firefox, enter the GitHub official URL, and press Enter to confirm normal network connection.
2. Step 2: Search OpenClaw Project: Type openclaw in the GitHub search bar, press Enter, and click the official openclaw/openclaw project.
3. Step 3: Select Stable Release Version: Find the Releases section on the project page, click the latest Release (2026.3.13) to enter the details page.
4. Step 4: Download Source Code ZIP: Locate Source code at the bottom of the version page, click to download the ZIP package, and wait for completion.
5. Step 5: Extract and Move Files: Right-click the ZIP package in the download directory, select Extract, then move the extracted folder to the pre-created work directory.

3. Build Dependency Environment: Install pnpm Package Manager

OpenClaw compilation and installation rely on the pnpm package manager, which is not pre-installed in the system and needs global deployment via npm. This is a key link connecting source code and compilation, requiring the administrator password during installation; version verification is needed after installation to ensure the tool works properly.

Some users may encounter environment variable invalidation after installation, which can be fixed by executing the source command to refresh configurations, ensuring the pnpm command can be called globally.

1. Step 1: Open Terminal Window: Right-click the Ubuntu desktop and select Open in Terminal to launch the command line tool.
2. Step 2: Install pnpm Globally: Run sudo npm install -g pnpm in the terminal, enter the Ubuntu administrator password after pressing Enter, and wait for installation.
3. Step 3: Verify pnpm Installation: After installation, run npm -v and pnpm -v in sequence to check versions (npm ≥ 10.9.4, pnpm ≥ 10.32.1).
4. Step 4: Fix Global Path Error (Optional): If a link error occurs later, run pnpm setup first, then execute source /home/hello/.bashrc to refresh environment variables.

4. Core Compilation & Installation: Finish OpenClaw Deployment in 3 Steps

This is the core part of OpenClaw deployment, completing dependency installation, UI construction and project compilation through three pnpm commands in sequence. Check the terminal feedback after each command to ensure no errors before proceeding to the next step.

After compilation, register OpenClaw as a global CLI command for direct terminal invocation. Solutions to possible path errors during registration are provided in the tutorial to ensure the command takes effect globally.

1. Step 1: Navigate to Source Directory: Use the cd command in the terminal to switch to the OpenClaw source folder in the work directory; copy the path to avoid input errors.
2. Step 2: Install Project Dependencies: Run pnpm install and wait for the terminal to automatically download and install all packages; proceed when completion prompt appears.
3. Step 3: Build UI Interface: Execute pnpm ui:build after dependency installation; this command runs fast, wait for the completion feedback.
4. Step 4: Compile Project Source Code: Run pnpm build and wait for the compilation to finish; the main OpenClaw program is now installed.
5. Step 5: Register Global Command: Execute pnpm link –global to register OpenClaw as a global CLI; re-run after fixing errors if needed.

5. Cross-End Connection: Link OpenClaw with Windows Ollama LLM

OpenClaw requires a large language model to implement AI functions. This tutorial uses local Windows Ollama to provide model support, requiring three core operations: network configuration, URL filling and model selection. This step is critical for normal OpenClaw interaction, so parameters must be checked carefully.

First, enable Ollama network exposure to allow the Ubuntu virtual machine to access the Windows Ollama service. Then fill in the correct Ollama access URL and select the appropriate operating mode. Finally, complete cloud login and model selection, skipping unnecessary plug-ins and API configurations.

1. Step 1: Launch OpenClaw Wizard: Run openclaw onboard –install-daemon in the terminal to start the configuration wizard and install the background daemon.
2. Step 2: Confirm Basic Settings: Use arrow keys to select Yes to continue, choose QuickStart mode and press Enter.
3. Step 3: Select Ollama as Model Source: Choose Ollama on the model selection interface and press Enter to enter URL configuration.
4. Step 4: Configure Ollama URL: Enter the Windows Ollama address (format: http://host-ip:11434) and press Enter; query the host IP via ipconfig on Windows.
5. Step 5: Choose Ollama Mode: Select Cloud + Local mode, copy the login link, open it in a browser to complete Ollama cloud connection.
6. Step 6: Finish Subsequent Config: Select Yes in the terminal to confirm login, use the default Kimi 2.5 model, skip chat channels, API keys and unused skills.

6. Final Verification: Ensure OpenClaw Runs Normally

After all configurations, verify that OpenClaw starts normally, the Web UI is accessible, and AI interaction is smooth — this is the final criterion for judging deployment success. It is recommended to save core access links for quick management interface access later.

If the Web UI opens properly and the AI assistant responds normally, the deployment is fully successful. OpenClaw will run in the background as a daemon process and start automatically on boot without manual activation.

1. Step 1: Select Interaction Mode: Choose Open the Web UI on the final configuration interface and press Enter to launch the web interface.
2. Step 2: Access Control Console: Copy the Control UI link from the terminal and paste it in Firefox to open the OpenClaw management panel.
3. Step 3: Test AI Interaction: Input simple commands (e.g., “Hello”, “What can you do”) in the console to check if the AI responds normally.
4. Step 4: Save Key Information: Archive the Control UI link, Ollama address and other important details for future maintenance and reconfiguration

7. Demo Video

You can watch the following demo video by select the subtitle to your preferred subtitle language.