Software Installation Instruction

Author

Kwok-leong Tang

Published

April 1, 2026

Pre-Workshop Instructions

This guide helps you install the software we will use during the morning session of the workshop. Please set aside 20–30 minutes, depending on your network speed, to complete the installation before the workshop.

Prerequisites

  • Supported OS: macOS (Apple Silicon M1/M2/M3/M4) or Windows 10/11
  • Disk space: At least 2–5 GB free (the model requires additional space)
  • RAM: 4 GB or more free RAM recommended
  • Internet access to download the application and the model
  • Admin rights to install applications
Important

Intel Mac users: LM Studio does not support Intel-based Macs. If your Mac has an Intel processor, please install Ollama instead (see Alternative: Ollama below).

To check your chip: click the Apple menu () → About This Mac. If it says “Chip: Apple M1/M2/M3/M4”, you can use LM Studio. If it says “Processor: Intel”, please use Ollama.

LM Studio

LM Studio is a free desktop application that lets you run large language models (LLMs) locally and privately on your own computer. It provides a chat interface so you can experiment with AI models without sending data to the cloud.

Part 1: Download LM Studio

  1. Open your web browser and go to https://lmstudio.ai.

  2. Click the Download button. The site automatically detects your operating system.

  3. Alternatively, go directly to https://lmstudio.ai/download and select your configuration:

    • macOS: Apple Silicon (M series) — requires macOS 13.4 or newer
    • Windows: x64 (requires AVX2 support) or ARM (Snapdragon X Elite)
    • Linux: x64, distributed as an AppImage (Ubuntu 20.04+)

  4. Click the green “Download” button to start downloading the installer.

Part 2: Install LM Studio

macOS (Apple Silicon only — M1/M2/M3/M4)

  1. Once the .dmg file has finished downloading, double-click it to open the disk image.
  2. Drag the LM Studio icon into the Applications folder.
  3. Open LM Studio from your Applications folder or via Spotlight.
  4. If macOS displays a security warning (Gatekeeper), go to System Settings → Privacy & Security and click “Open Anyway”.
Note

For MLX model support (which provides better performance on Apple Silicon), macOS 14.0 or newer is required.

Windows 10/11

  1. Once the .exe installer has finished downloading, double-click it to run.
  2. Follow the on-screen installation prompts (accept the license agreement, choose an install location, etc.).
  3. If Windows SmartScreen appears, click More info → Run anyway.
  4. After installation completes, launch LM Studio from the Start menu or desktop shortcut.

Part 3: Download the Qwen3.5-0.8B Model

We will use the Qwen3.5-0.8B model for in-class practice. This is a small but capable model (~1 GB) that runs comfortably on most machines.

  1. Open LM Studio on your computer.

  2. Click the magnifying glass icon (Search / Discover) in the left sidebar to open the model search page.

  3. In the search bar, type:

    Qwen3.5-0.8B

  4. The search results will show available versions:

    • macOS users: Look for the MLX format (e.g., Qwen3.5-0.8B-MLX) for best performance on Apple Silicon.
    • Windows users: Look for the GGUF format (e.g., Qwen3.5-0.8B-GGUF).

  5. For the GGUF format, you may see different quantizations. Choose Q4_K_M for a good balance of quality and size.

  6. Click the Download button next to your chosen version. The model is relatively small (typically under 1 GB).

  7. Once the download is complete, you can load the model by selecting it from the model dropdown at the top of the chat interface.

Part 4: Verify Your Installation

  1. Open LM Studio and load the Qwen3.5-0.8B model.

  2. In the chat interface, send a simple prompt:

    Hello! What is 2 + 2?

  3. Confirm you get a response. This verifies the model is running locally on your machine.

Tip

If you have extra time and disk space, you may also download a larger model such as Qwen3.5-4B for higher-quality responses. Expect each model to require several GBs of storage.

Alternative: Ollama

If you cannot run LM Studio (e.g., Intel Mac, or you prefer a command-line tool), Ollama is an excellent alternative that also runs models locally.

Install Ollama

macOS (Intel and Apple Silicon)

  1. Go to https://ollama.com/download.
  2. Download the macOS installer and follow the prompts.

Windows

  1. Go to https://ollama.com/download.
  2. Download the Windows installer and follow the prompts.

Download the Qwen3.5-0.8B Model

After installing Ollama, open your Terminal (macOS) or Command Prompt / PowerShell (Windows) and run:

ollama run qwen3.5:0.8b

This will automatically download (~1 GB) and start the model. You can begin chatting with it directly in the terminal.

Verify Ollama Installation

After the model loads, type a simple prompt in the terminal:

Hello! What is 2 + 2?

If you get a response, your installation is working correctly. You can exit by typing /bye.

Troubleshooting

  • macOS Gatekeeper warning: Go to System Settings → Privacy & Security → Open Anyway after the first launch attempt.
  • Windows SmartScreen: Click More info → Run anyway if downloaded from the official site.
  • Slow performance: Close other memory-intensive applications. Smaller models like the 0.8B run comfortably on CPU; no GPU is required.
  • Model not loading: Ensure you have enough free disk space and RAM. Try restarting LM Studio or Ollama.
  • Network/firewall prompts: Allow the application to access the network so it can download models.

Google Antigravity

Google Antigravity is a free AI-powered development platform from Google that functions as an agentic IDE — it treats AI as an autonomous actor capable of planning, executing, validating, and iterating on complex engineering tasks with minimal human intervention.

Prerequisites

  • Supported OS: macOS 10.15 (Catalina) or later (Intel and Apple Silicon), Windows 10 or later
  • Disk space: At least 500 MB free
  • RAM: 4 GB minimum (8 GB recommended)
  • Internet access with at least 5 Mbps recommended
  • Google Account: A personal Gmail account is required (Google Workspace accounts are not currently supported)
  • Chrome web browser installed

Part 1: Download Google Antigravity

  1. Open your web browser and go to https://antigravity.google/download.

  2. Select your operating system:

    • macOS: Download the .dmg installer (supports both Intel and Apple Silicon)
    • Windows: Download the .exe installer

  3. The download is approximately 300–400 MB and typically takes 5–10 minutes depending on your connection speed.

Part 2: Install Google Antigravity

macOS

  1. Once the .dmg file has finished downloading, double-click it to open the disk image.
  2. Drag the Antigravity icon into the Applications folder.
  3. Open Antigravity from your Applications folder or via Spotlight.
  4. On first launch, macOS may ask you to grant permissions in System Settings → Privacy & Security. Click “Open Anyway” if prompted.

Windows 10/11

  1. Once the .exe installer has finished downloading, double-click it to run.
  2. If Windows SmartScreen appears, click More info → Run anyway.
  3. Follow the on-screen installation prompts.

Part 3: Initial Setup

  1. When Antigravity launches for the first time, you can choose to import settings from VS Code/Cursor or start fresh.

  2. Select your preferred editor theme (dark or light).

  3. Configure agent execution policies (you can keep the defaults for now):

    • Terminal policy
    • Review policy
    • JavaScript policy

  4. Configure editor settings (keybindings, extensions, command line tools).

  5. Sign in with your personal Gmail account. You will be redirected to Chrome to authenticate, then returned to Antigravity.

  6. Accept the Terms of Use.

Part 4: Verify Your Installation

  1. After signing in, you should see the Agent Manager — the main workspace for dispatching and monitoring autonomous agents.

  2. Try creating a simple task by typing a prompt such as:

    Create a Python script that prints "Hello, World!"

  3. Confirm that the agent responds and generates code. This verifies that Antigravity is connected to your Google account and working correctly.

Note

Antigravity uses Google’s Gemini models via the cloud. Unlike LM Studio and Ollama, it does not run models locally — an internet connection is always required.

Uninstall (if needed)

  • LM Studio on macOS: Drag the app from Applications to Trash. Optionally remove caches under ~/Library/Application Support/LM Studio/.
  • LM Studio on Windows: Use Settings → Apps to uninstall. Optionally remove leftover folders in %AppData% or %LocalAppData%.
  • Ollama on macOS: Run ollama rm qwen3.5:0.8b to remove the model. Delete the Ollama app from Applications.
  • Ollama on Windows: Uninstall via Settings → Apps.
  • Antigravity on macOS: Drag the app from Applications to Trash.
  • Antigravity on Windows: Uninstall via Settings → Apps.