Module 3: LM Studio - Introduction

Chat Interface and Playground

Overview

You'll learn to use LM Studio's graphical interface to chat with your local model without writing code.

Time: 20 minutes
Difficulty: Low


🎯 Objectives

  • ✅ Use the Chat interface
  • ✅ Adjust parameters (temperature, max tokens)
  • ✅ Create custom system prompts
  • ✅ Save conversations

💬 Chat Interface

Access:

  1. Model loaded (🟢 in the top bar)
  2. Click the 💬 Chat tab

System Prompt:

Defines the model's behavior:

You are a technical assistant specialized in Python.
Answer clearly and concisely.

Send messages:

Type in the input box and press Enter.


⚙️ Parameters Panel

Temperature (0.0-2.0):

  • 0.0: Deterministic
  • 0.7: Balanced (default)
  • 2.0: Very creative

Max Tokens:

  • Response limit
  • Default: 512
  • Range: 1-32768 (depending on the model)

Top-p (0.0-1.0):

  • Alternative to temperature
  • Default: 0.9

💾 Save Conversations

Export chat:

  • Click the "Export" button
  • Format: JSON or Markdown
  • Saves to disk

📊 Comparison with OpenAI

FeatureLM StudioOpenAI Playground
Cost$0$$$
Privacy100% localCloud
SpeedVariable (hardware)Consistent (~1.5s)
ModelsOpen sourceGPT-3.5/4

✅ Summary

  • A simple interface for chatting
  • Adjustable parameters (temperature, max_tokens)
  • Customizable system prompts
  • Conversation export

Next: 05-api-local.md