OpenClaw + LattePanda IOTA: Building a Home AI Robot Assistant

LattePanda IOTA is a compact platform well suited for building home AI and IoT systems, powered by an Intel N150 quad-core processor and configurable with up to 16 GB memory. Its small form factor and balanced performance make it ideal for always-on devices such as a home AI robot assistant.

Why choose LattePanda IOTA for an OpenClaw robot?

• Palm size: easy to fit inside a robot controller.
• Balanced performance: handles AI, vision, and control tasks well.
• High cost-performance ratio: especially valuable as memory and storage prices rise.
• Easy Ubuntu support: smooth x86 compatibility with minimal driver issues.
• Rich expansion ( UPS Exp. / 4G LTE Exp. / M.2 M-Key Exp. / PoE Exp.): supports cameras, displays, storage, and networking.
• Low-power passive cooling: silent and reliable for long-term operation.
• The built-in RP2040: simplifies real-time control and sensor integration.

Hardware Assembly

For the first boot, we recommend preparing the following hardware to make setup and debugging easier:

• USB keyboard and mouse
• Standard HDMI cable
• A display with HDMI input
• USB Type-C power adapter with 15V PD output
• a PH2.0 4-pin power connector (10–15V DC, 15V max)

⚠️ Note: The power supply must provide at least 24W to ensure stable operation.

LattePanda IOTA prioritizes power input from the USB Type-C port. If the USB-C power is disconnected, the system will automatically and silently switch to the PH2.0 power input. When USB-C power is restored, it switches back without rebooting or data loss. This failover mechanism ensures uninterrupted operation during power transitions.

Power Requirements

• Supports USB Power Delivery (PD)
• Must negotiate a 15V output
• Minimum power requirement: 15V @ 1.8A

Most standard USB Type-C PD power adapters or power banks available on the market meet these requirements.

The power connection method is shown below.

For the first boot, it is recommended to prepare a USB flash drive of 8 GB or larger and use Rufus to create a bootable Ubuntu installer.

Download: https://releases.ubuntu.com/jammy/

Download the Ubuntu image from the link above and write it to your 8 GB USB flash drive.

After creating the bootable USB drive, plug it into any USB 3.2 port on the LattePanda IOTA. Power on the device and repeatedly press F7 during startup until the Boot Device Selection Menu appears, then select the USB drive to boot.

You will then enter the standard Ubuntu installer. On the “Try or Install Ubuntu” screen, select Install Ubuntu.

Follow the on-screen steps to choose the language and keyboard layout, then select the disk option. Make sure to choose Full disk install to use the entire storage. After that, set your time zone and enter your name, computer name, username, and password.

Once the installation is complete, reboot into Ubuntu. After logging in, open a terminal. In this setup, we first enable OpenSSH Server to allow remote access to the LattePanda IOTA via SSH.

Deploying OpenClaw

Start by updating the system and installing basic dependencies in the terminal:

 Install Node.js:

After the installation completes, verify the versions:

Next, install pnpm:

Finally, install OpenClaw:

After the installation is complete, run the following command:

Follow the on-screen instructions to complete the configuration.

Use the left and right arrow keys to navigate through the options.

Select Quick Start.

Select the API provider for the large language model you are using.

In this setup, we selected Moonshot AI (Kimi 2) as the language model provider.

We selected the Kimi Code API Key, which can be obtained by subscribing at: https://kimi.com/code/

Create a new API key in the dashboard.

Select Create New. 

Copy your API key and store it securely, as it will only be shown once.

Paste the API key back into the terminal when prompted.

Keep the default model selection.

Next, select the communication channel.

Proceed to install the plugins.

If you encounter the following issue:

You can resolve this by reinstalling it later.

Since OpenClaw is updated frequently, the Feishu plugin may occasionally encounter issues during updates. You can skip it for now.

When prompted to configure skills, we recommend skipping this step. Most of the available skills are macOS-specific and are not required for this setup.

If you want to explore the options, select Yes to take a look first.

Since these require installation via Homebrew, you can safely skip them.

Choose the installation method based on your preference. In this setup, we used npm by default, but you can also try pnpm if preferred.

You can select “Skip for now”, as most of these options are optional and some require paid services. The remaining steps also prompt for API keys—if you don’t have them, simply choose No.

You will see several informational prompts—feel free to review them briefly. On the first startup, the gateway service may report an error.

You can access the control UI directly via the highlighted link shown in the terminal. Open this URL in a browser on the desktop.

Then copy the TOKEN value and paste it into the corresponding field in the UI, following the steps shown in the image below.

Complete the initialization process to finish the setup.

After the setup completes, make sure to start the OpenClaw gateway, as this service is a core component of the system.

You can also open a terminal and run the following command to adjust web-related settings:

If you have a Brave Search API key, you can add it here.

Continue until the process finishes. You can then safely stop the other terminal session.

Start the OpenClaw gateway and enable detailed output:

If the startup fails, stop the service and try again. You can run:

Then start the gateway once more.

After the gateway starts successfully, you can re-enter the configuration using:

Navigate to the Feishu configuration section to set up the Feishu bot. Before doing so, make sure to log in to the Feishu developer console.

Follow the steps below to create an application:

Select Bot.

Configure the bot’s permissions, enable events and callbacks, then publish it to make it operational.

Enable the required permissions.

We briefly configured the IM-related permissions. Enter im:message to check which features are available.

We selected send_as_bot.

Click on Credentials & Basic Information.

Copy both pieces of information and register them in the terminal.

Copy the ID and Secret.

Paste them into the terminal and press Enter.

Enter the TUI to see the results.

Previously, we named the bot XiaoTianTian, and it started successfully.

Next, configure the Feishu bot by setting up events and callbacks and selecting long-polling as the subscription method.

Then configure the callback settings.

I activated the Feishu plugin and restarted the gateway. To view the startup logs, I added the --verbose flag and confirmed that the Feishu plugin had started successfully.

Then, in the Feishu app, add the events you want to subscribe to.

Confirm the callback information and use long-polling for the connection.

Finally, publish the application by creating a new version in the console. Only after publishing will it be usable.

Then, simply save the settings.

A new window will appear. Open the application to see it in action.

The first interaction will prompt you to approve pairing with Feishu. Execute the command indicated by the arrow.

The bot can now run continuously. After logging in with Feishu on my phone, the robot is ready to work at home.

Also, be mindful of the number of tokens used by your large language model API.