Skip to content

🎮 LM Studio Demo Agent

The LM Studio Demo Agent showcases GUI-based model management with LM Studio in SuperOptiX. This demo focuses specifically on how to configure and use LM Studio models with a visual interface for model management.

🎯 What This Demo Shows

This demo demonstrates:

  • 🎮 LM Studio Model Integration: How to configure LM Studio models in SuperOptiX
  • 🖥️ GUI Model Management: Visual interface for model management
  • 🏠 Local Model Usage: Running models completely offline
  • ⚙️ Playbook Configuration: How to set up LM Studio in agent playbooks

🚀 Setup LM Studio Model

1. Install LM Studio

Bash
# Download and install LM Studio from https://lmstudio.ai
# Launch LM Studio and download a model through the interface

2. Install LM Studio Model

Bash
# Install the LM Studio model used in this demo
super model install -b lmstudio llama-3.2-8b-instruct

3. Start LM Studio Server

Bash
# Start LM Studio server on port 1234
super model server lmstudio llama-3.2-8b-instruct --port 1234

4. Pull and Run the Demo

Bash
# Pull the LM Studio demo agent
super agent pull lmstudio_demo

# Compile the agent
super agent compile lmstudio_demo

# Run the agent
super agent run lmstudio_demo --goal "What are the key features of LM Studio?"

🔧 LM Studio Configuration in Playbook

The LM Studio demo showcases how to configure LM Studio models in the agent playbook:

Language Model Configuration

YAML
language_model:
  location: local
  provider: lmstudio
  model: llama-3.2-8b-instruct
  api_base: http://localhost:1234
  temperature: 0.7
  max_tokens: 2048

Key Configuration Points:

  • 🎯 provider: lmstudio: Specifies LM Studio as the model backend
  • 🤖 model: The LM Studio model identifier
  • 🌐 api_base: LM Studio server endpoint (default: http://localhost:1234)
  • 🌡️ temperature: Controls response creativity (0.7 = balanced)
  • 📏 max_tokens: Maximum response length

🎮 LM Studio: Visual AI Management

LM Studio brings the power of local AI with the simplicity of a graphical interface. Perfect for users who prefer visual tools:

  • 🖥️ Visual Interface: Beautiful GUI for managing models and conversations
  • 📊 Real-time Monitoring: Watch your model's performance in real-time
  • 🎯 Easy Model Selection: Browse and select models with a visual interface
  • 🖱️ Point-and-Click: No command line required for basic operations
  • 🪟 Windows Native: Optimized for Windows users with familiar interface
  • 🍎 macOS Support: Also works great on macOS systems

🔧 Customizing LM Studio Configuration

Change Model

Edit agents/lmstudio_demo/playbook/lmstudio_demo_playbook.yaml:

YAML
language_model:
  model: llama-3.2-1b-instruct  # Different LM Studio model
  api_base: http://localhost:1234

Adjust Performance Settings

YAML
language_model:
  temperature: 0.5  # More precise responses
  max_tokens: 4096  # Longer responses

Use Different Port

YAML
language_model:
  api_base: http://localhost:8080  # Custom port

🚨 Troubleshooting LM Studio

Common Issues

  1. LM Studio Server Not Running

    Bash
    # Check if LM Studio server is running
curl http://localhost:1234/v1/models

# Start LM Studio server
super model server lmstudio llama3.2:8b --port 1234

  2. Model Not Installed

    Bash
    # Check installed LM Studio models
super model list --backend lmstudio

# Install the required model
super model install -b lmstudio llama3.2:8b

  3. Performance Issues

  4. Ensure sufficient RAM (8GB+ recommended)
  5. Close other resource-intensive applications
  6. Consider using smaller models for faster responses

Getting Help

Bash
# Check agent status
super agent inspect lmstudio_demo

# View agent logs
super agent logs lmstudio_demo

# Get LM Studio help
super model server --help

🔗 Next Steps

  1. Try Other Model Backends: Explore MLX, Ollama, or HuggingFace demos

Ready to explore GUI model management? Start with the LM Studio demo! 🚀