How to Run OpenClaw on Raspberry Pi: Complete 2026 Setup Guide

Running OpenClaw on a Raspberry Pi is one of the most cost-effective ways to have a 24/7 personal AI assistant. For about $1/month in electricity, your Pi becomes an always-on agent that manages your calendar, answers messages, runs automations, and keeps your data entirely private.

This guide walks you through every step: choosing the right hardware, installing Raspberry Pi OS, setting up Node.js, deploying OpenClaw, connecting Telegram, and optimizing performance. We also cover the real limitations you should know about before committing to this path.

---

Why Raspberry Pi for OpenClaw?

OpenClaw is an open-source AI assistant and one of the most popular open-source AI projects. It connects to Telegram, WhatsApp, Discord, and Slack, and runs agentic workflows: scheduling, email, web browsing, custom skills, and proactive notifications.

Unlike cloud-hosted chatbots, OpenClaw needs to run 24/7 to be useful. It sends heartbeat checks, maintains persistent memory, and monitors your tasks around the clock. That makes a Raspberry Pi surprisingly ideal:

• Ultra-low power: ~5W average, costing $0.50-1.00/month in electricity
• Always on: No laptop to close, no desktop to restart
• Complete privacy: Your API keys, conversations, and business data never leave your home network
• Compact: Fits on a shelf, silent operation with passive cooling
• Affordable: $60-100 for the hardware, then near-zero running costs

The tradeoff? You need to be comfortable with the command line, and you accept responsibility for maintenance, updates, and security. If that sounds fine, let's get started.

---

Hardware Requirements: Which Pi Do You Need?

Not all Raspberry Pi models are equal for running OpenClaw. Here's what actually matters:

Raspberry Pi Model Comparison

Raspberry Pi 5 (8GB) - Recommended

• 2.4GHz quad-core Arm Cortex-A76
• 8GB LPDDR4X-4400 RAM
• PCIe 2.0 for NVMe SSD
• Best performance for multi-skill workflows
• Price: ~$80

Raspberry Pi 5 (4GB) - Good

• Same CPU as above
• 4GB RAM limits concurrent skills
• Fine for single-agent, single-channel setups
• Price: ~$60

Raspberry Pi 4 (8GB) - Acceptable

• 1.8GHz quad-core Cortex-A72
• 8GB LPDDR4-3200 RAM
• No native PCIe (USB 3.0 for external SSD)
• Works well but noticeably slower on heavy tasks
• Price: ~$55-65 (often discounted)

Raspberry Pi 4 (4GB) - Minimum Viable

• Same CPU as Pi 4 8GB
• Tight on memory with multiple skills running
• Expect swap usage under load
• Only recommended for testing or very light workloads

Why 8GB RAM Matters

OpenClaw itself uses modest resources, but the Node.js runtime, message processing, skill execution, and persistent memory all add up. With 4GB, you will hit swap regularly when running multiple skills or handling conversation threads with large context windows. That means SD card wear and slower responses.

With 8GB, you have comfortable headroom for:

• Running OpenClaw with 3-5 active skills
• Handling multiple simultaneous conversations
• Running local tools alongside OpenClaw (monitoring, backup scripts)
• Future-proofing as OpenClaw adds more features

Full Shopping List

Here's what you need to buy:

Important: Do not cheap out on the power supply. Under-voltage causes random crashes during AI workloads. The official Raspberry Pi power supply is specifically rated for sustained high loads. Generic USB-C chargers from Amazon often cannot deliver stable 5A output.

---

Step 1: Install Raspberry Pi OS

We will use Raspberry Pi OS Lite (64-bit) since we don't need a desktop environment. This saves RAM and CPU for OpenClaw.

Flash the OS

1. Download Raspberry Pi Imager on your computer

2. Insert your microSD card (or NVMe SSD via USB adapter)

3. Select Raspberry Pi OS Lite (64-bit) - Debian Trixie based

4. Click the gear icon to pre-configure:

   • Set hostname (e.g., openclaw-pi)
   • Enable SSH with password authentication
   • Set username and password
   • Configure your WiFi network (or use Ethernet)
   • Set your locale and timezone

5. Flash the image and insert the card/SSD into your Pi

First Boot

Power up and SSH in:

ssh youruser@openclaw-pi.local

Update everything first:

sudo apt-get update && sudo apt-get upgrade -y

Configure locale and timezone if you skipped it earlier:

sudo raspi-config

Navigate to Localisation Options and set your locale, timezone, and keyboard layout.

Optional: Boot from NVMe SSD

If you bought an NVMe SSD and M.2 HAT+:

1. Install the HAT and SSD on your Pi 5
2. Boot from the SD card first
3. Use SD Card Copier (under Accessories) to clone to the NVMe
4. Update the boot order in raspi-config > Advanced Options > Boot Order > NVMe

This gives you 3-5x faster read/write speeds and eliminates SD card corruption risks. For a 24/7 server, this is highly recommended.

---

Step 2: Install Node.js

OpenClaw requires Node.js 22.x or later. The version in Raspberry Pi OS repos is usually outdated, so we install via NodeSource:

# Install Node.js 22.x via NodeSource
curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash -
sudo apt-get install -y nodejs

# Verify
node --version  # Should show v22.x.x
npm --version

Alternatively, if you prefer version management with fnm:

curl -fsSL https://fnm.vercel.app/install | bash
source ~/.bashrc
fnm install 22
fnm use 22

Install essential build tools (some OpenClaw skills need native compilation):

sudo apt-get install -y build-essential git python3

---

Step 3: Install OpenClaw

Now for the main event. OpenClaw installation is straightforward:

# Install OpenClaw globally
npm install -g openclaw

# Verify installation
openclaw --version

Initial Configuration

Run the setup wizard:

openclaw onboard

This walks you through:

• AI provider: Choose your LLM provider (Anthropic, OpenAI, etc.)
• API key: Enter your API key for the chosen provider
• Agent name: Name your assistant
• Default model: Select which model to use (claude-opus-4-6, gpt-4.1, etc.)

The configuration is saved to ~/.openclaw/openclaw.json.

Set Up the Gateway

The OpenClaw gateway is the always-on daemon that keeps your assistant running:

# Start the gateway
openclaw gateway start

# Check status
openclaw gateway status

# View logs
openclaw gateway logs

To make OpenClaw start automatically on boot:

# Create a systemd service
sudo tee /etc/systemd/system/openclaw.service > /dev/null << 'EOF'
[Unit]
Description=OpenClaw AI Assistant Gateway
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
User=youruser
WorkingDirectory=/home/youruser
ExecStart=/usr/bin/openclaw gateway start --foreground
Restart=always
RestartSec=10
Environment=NODE_ENV=production

[Install]
WantedBy=multi-user.target
EOF

sudo systemctl daemon-reload
sudo systemctl enable openclaw
sudo systemctl start openclaw

Now OpenClaw survives reboots and restarts automatically if it crashes.

---

Step 4: Connect Telegram

Telegram is the most popular channel for OpenClaw on Raspberry Pi. Here's how to set it up:

Create a Telegram Bot

1. Open Telegram and search for @BotFather
2. Send /newbot
3. Choose a name (e.g., "My AI Assistant")
4. Choose a username (must end in bot, e.g., myai_assistant_bot)
5. Copy the API token BotFather gives you

Configure OpenClaw for Telegram

openclaw channel add telegram

When prompted, paste your bot token. OpenClaw will verify the connection and start listening for messages.

Test It

Send a message to your bot on Telegram. You should see a response within a few seconds. Try:

• "What can you do?"
• "What time is it?"
• "Remind me to check email in 30 minutes"

WhatsApp Alternative

If you prefer WhatsApp, OpenClaw also supports it via the Baileys library (no Business API needed). The setup is more complex and requires a dedicated phone number. For most self-hosters, Telegram is the simplest starting point.

---

Step 5: Performance Optimization

A Raspberry Pi is not a Mac Studio. Here are practical tips to get the best performance:

Use an SSD

This is the single biggest performance improvement. SD cards have limited write endurance and slow random I/O. OpenClaw writes to memory files, logs, and conversation history constantly. An NVMe SSD makes everything feel snappier and prevents premature SD card failure.

Reduce Swap Usage

With 8GB RAM, you may not need much swap. But if you're on 4GB:

# Check current swap
free -h

# Increase swap if needed (for 4GB Pi)
sudo dphys-swapfile swapoff
sudo sed -i 's/CONF_SWAPSIZE=.*/CONF_SWAPSIZE=2048/' /etc/dphys-swapfile
sudo dphys-swapfile setup
sudo dphys-swapfile swapon

Overclock (Pi 5 Only)

The Pi 5 can be safely overclocked for better performance:

# Edit config
sudo nano /boot/firmware/config.txt

# Add these lines
arm_freq=2800
gpu_freq=1000
over_voltage_delta=50000

Only do this with adequate cooling (active fan or heatsink case). Monitor temperatures with:

vcgencmd measure_temp

Keep it below 80C under sustained load.

Optimize Node.js Memory

Limit Node.js heap size to prevent out-of-memory kills:

# Add to your .bashrc or systemd service
export NODE_OPTIONS="--max-old-space-size=2048"

For 4GB Pi, use --max-old-space-size=1024.

Disable Unnecessary Services

Free up resources by disabling services you don't need:

# Disable Bluetooth if not using it
sudo systemctl disable bluetooth
sudo systemctl stop bluetooth

# Disable audio (headless server doesn't need it)
sudo systemctl disable pulseaudio 2>/dev/null

---

Step 6: Monitoring and Maintenance

Running a 24/7 server means monitoring it. Here's a minimal but effective setup:

Health Check Script

Create a simple script that checks if OpenClaw is running:

#!/bin/bash
# /home/youruser/check-openclaw.sh

if ! openclaw gateway status | grep -q "running"; then
    echo "$(date): OpenClaw is down, restarting..." >> /var/log/openclaw-health.log
    openclaw gateway restart
fi

Add it to cron:

crontab -e
# Add this line (checks every 5 minutes)
*/5 * * * * /home/youruser/check-openclaw.sh

Monitor Resources

Keep an eye on memory and CPU:

# Quick check
htop

# Temperature
vcgencmd measure_temp

# Disk usage
df -h

# Memory
free -h

Automatic Updates

Keep your system secure with unattended upgrades:

sudo apt-get install unattended-upgrades
sudo dpkg-reconfigure -plow unattended-upgrades

For OpenClaw updates, check periodically:

npm outdated -g openclaw
npm update -g openclaw

Warning: Always check the OpenClaw changelog before updating. Breaking changes can affect your skills and configuration.

Backup Your Configuration

Your OpenClaw setup lives in ~/.openclaw/. Back it up regularly:

# Simple backup to USB or network share
tar -czf ~/openclaw-backup-$(date +%Y%m%d).tar.gz ~/.openclaw/

---

Real Limitations You Should Know

Let's be honest about what a Raspberry Pi can and cannot do with OpenClaw:

What Works Well

• Single-user Telegram/Discord bot: Perfectly responsive for one person
• Scheduled tasks and automations: Heartbeats, reminders, monitoring
• Text-based skills: Email, calendar, web search, file management
• Always-on availability: The Pi excels at being a quiet, persistent server

What Struggles

• Local LLM inference: Running Ollama with even a 3B model is painfully slow on a Pi. You will want to use cloud APIs (Anthropic, OpenAI) for the actual AI processing
• Multiple simultaneous users: More than 2-3 concurrent conversations will cause latency spikes
• Heavy file processing: Image generation, video processing, or large document analysis will bottleneck quickly
• Browser automation: Running Puppeteer or Playwright consumes significant RAM

The Hidden Costs

The Raspberry Pi hardware is cheap, but the total cost of self-hosting includes:

• API costs: $20-100+/month depending on usage (Claude, GPT-4.1 tokens)
• Your time: Initial setup (2-4 hours), ongoing maintenance (1-2 hours/month), troubleshooting when things break
• Domain + DDNS: If you want external access ($10-15/year)
• Security responsibility: You are responsible for securing API keys, SSH access, and network exposure

A realistic monthly cost for a Raspberry Pi OpenClaw setup: $25-80/month (mostly API costs) plus your time.

---

When a Managed Service Makes More Sense

Self-hosting OpenClaw on a Raspberry Pi is a fantastic learning experience and gives you full control. But it's not for everyone.

If you are a business owner, coach, freelancer, or solopreneur who needs an AI assistant for customer service, appointment booking, or lead generation, consider whether spending hours on Linux configuration and Node.js debugging is the best use of your time.

ClawRapid deploys a fully managed OpenClaw instance in 60 seconds. No Raspberry Pi, no SSH, no systemd services, no security patches. You choose your AI model, connect your Telegram or WhatsApp, and start chatting with your customers.

What you get with ClawRapid for $45/month:

• Managed hosting: 99.9% uptime, automatic updates, zero maintenance
• **$15/mo AI credit included
• Security handled: Encrypted storage, managed SSL, regular security audits
• Multi-channel: Telegram, WhatsApp, Discord, Slack, all pre-configured
• Support: Real humans who help you configure your assistant for your specific business

Compare that to self-hosting: $85-150 hardware upfront, $25-80/month in API costs, 2-4 hours setup, 1-2 hours/month maintenance, and the stress of knowing that if your SD card corrupts at 2 AM, your business assistant goes offline.

Want to skip the setup and start helping customers today? Try ClawRapid free for 7 days - 60-second deployment, no credit card required.

---

FAQ

What is the minimum Raspberry Pi model for OpenClaw?

A Raspberry Pi 4 with 4GB RAM is the absolute minimum. We recommend a Pi 5 with 8GB for comfortable, reliable 24/7 operation. Older models (Pi 3 and below) do not have enough RAM or CPU power.

Can I run local AI models on a Raspberry Pi with OpenClaw?

Technically yes, via Ollama or llama.cpp. But inference is extremely slow - even a 3B parameter model takes 10-30 seconds per response on a Pi 5. For practical use, connect OpenClaw to cloud APIs (Anthropic Claude, OpenAI GPT) and use the Pi only as the agent runtime.

How much does it cost per month to run OpenClaw on Raspberry Pi?

Electricity costs about $0.50-1.00/month. But the real cost is API tokens: expect $20-100+/month depending on your model and usage. Total realistic cost: $25-80/month plus your time for maintenance.

Is it safe to expose my Raspberry Pi to the internet?

For most users, you do not need to. OpenClaw connects outbound to Telegram/WhatsApp APIs - it does not need an incoming port open. Avoid exposing SSH or the OpenClaw gateway to the public internet unless you know exactly what you are doing (use VPN or Tailscale instead).

Can I run OpenClaw on Raspberry Pi without a monitor?

Yes, and that is the recommended approach. Use Raspberry Pi OS Lite (no desktop) and manage everything via SSH. This saves significant RAM and CPU resources.

How do I update OpenClaw on Raspberry Pi?

Run npm update -g openclaw and restart the gateway with openclaw gateway restart. Always check the changelog first for breaking changes.

What happens if my Raspberry Pi loses power?

If you set up the systemd service as described above, OpenClaw will restart automatically when power returns. Your conversation history and memory are persisted to disk, so nothing is lost. Consider a small UPS ($20-30) for critical deployments.

---

Wrapping Up

Running OpenClaw on a Raspberry Pi is a genuinely practical option for developers and tinkerers who want full control over their AI assistant. The combination of low power consumption, always-on availability, and complete data privacy makes it an attractive self-hosting platform.

The key recommendations: get a Pi 5 with 8GB RAM, use an NVMe SSD instead of an SD card, connect to cloud AI APIs rather than running local models, and set up proper monitoring with systemd and cron.

If you are technical and enjoy managing infrastructure, a Raspberry Pi is one of the most cost-effective ways to run OpenClaw 24/7.

If you are a business owner who just wants an AI assistant that works, without worrying about power supplies, SSH keys, and Node.js versions, ClawRapid handles all of that for $45/month. Deploy in 60 seconds and focus on your business instead.

Ready to deploy your AI assistant? Try ClawRapid free for 7 days - $15 AI credit included, no hardware required.