LM Studio Integration (Raspberry Pi)

April 16, 2026 · View on GitHub

This example shows how to create a decentralized AI chatbot using QuIXI and LM Studio connected to the Ixian Platform via QuIXI.

The device receives Ixian chat messages, forwards them to a local LLM, and replies with AI-generated responses while still supporting basic system commands.

🛠 Requirements

  • Raspberry Pi 2+ or Linux host
  • Internet connection
  • LM Studio running locally (default: http://localhost:1234/v1/chat/completions)
  • .NET 10
  • Mosquitto MQTT

⚡ Installation

sudo apt update
sudo apt install -y curl git jq mosquitto mosquitto-clients

# Install .NET 10
curl -sSL https://dot.net/v1/dotnet-install.sh | bash /dev/stdin --channel 10.0 --version latest --verbose
echo 'export DOTNET_ROOT=$HOME/.dotnet' >> ~/.bashrc
echo 'export PATH=$PATH:$DOTNET_ROOT:$DOTNET_ROOT/tools' >> ~/.bashrc
source ~/.bashrc

# Clone and build QuIXI + Ixian Core
git clone https://github.com/ixian-platform/Ixian-Core.git
git clone https://github.com/ixian-platform/QuIXI.git
dotnet build -c Release

📂 Project Structure

/LMStudio
 ├── start.sh              # Launches QuIXI and message handler
 ├── ixiMessageHandler.sh  # Handles chat commands & forwards messages to LM Studio
 ├── helpers.sh            # Helper functions for messaging & system control
 ├── ixian.cfg             # Example QuIXI configuration file
 ├── quixi.service         # systemd service file for auto-start

Note: quixi.service assumes start.sh and all helper scripts reside in /root/. Adjust paths if you move them elsewhere.

▶️ Running the Example

Copy example ixian.cfg file to Release directory and edit if required

cp ixian.cfg QuIXI/QuIXI/bin/Release/net10.0/

Manual start

chmod +x start.sh ixiMessageHandler.sh helpers.sh
./start.sh

Auto-start on boot (systemd)

Copy the provided service file:

sudo cp quixi.service /etc/systemd/system/
sudo systemctl enable quixi
sudo systemctl start quixi

Check status:

systemctl status quixi

🔑 Add Your Device as a Contact

Before you can send commands and messages to AI, add your Ixian address as a contact to QuIXI. Replace DEVICE_ADDRESS with the address of your Spixi client:

curl --get \
  --data-urlencode "address=DEVICE_ADDRESS" \
  "http://localhost:8001/addContact"

💡 Supported Commands

Send commands via Ixian chat (Chat/# topic):

CommandAction
tempReplies with Raspberry Pi CPU temperature
wifiManage Wi-Fi connections (add/remove/list)
contactsManage Ixian contacts (accept/add/remove/list)
helpShow available commands
Any otherForwarded to LM Studio for AI-generated reply

🔌 How It Works

  1. start.sh launches QuIXI and the message handler.

  2. ixiMessageHandler.sh listens to Chat/# and parses incoming messages:

    • Built-in commands (temp, wifi, contacts, help) handled locally.
    • All other messages forwarded to LM Studio via HTTP API.

  3. AI responses are sent back to the sender via Ixian chat.

This demonstrates how Ixian + QuIXI can be used to create a decentralized AI assistant.

📜 License

This example is part of QuIXI and licensed under the MIT License.