Skip to content

Quickstart Guide

Get Backend.AI GO up and running and start your first local AI conversation in 5 minutes.

Step 1: Install Backend.AI GO

Download the installer for your platform from the official website or the GitHub Releases page.

  • macOS: Open the .dmg file and drag Backend.AI GO to your Applications folder.
  • Windows: Run the .exe or .msi installer and follow the prompts.
  • Linux: Install the .deb package or use the .flatpak bundle.

For more details, see the Installation Guide.

Step 2: Complete First-Time Setup

When you first launch Backend.AI GO, you'll be guided through a setup wizard:

  1. Select language: Choose the interface language.
  2. Select recommended options: The app picks the inference engine that suits your computer. You can change it here.
  3. Select model location: Confirm or change where models are stored.
  4. Select first AI model (optional): Pick a recommended model to download right away.
  5. Connect cloud AI providers (optional): Add API keys for cloud models.
  6. All set: Enter the app.

For detailed information about each step, see First-Time Setup.

Step 3: Download Your First Model

After completing the initial setup, the app will start. You need to download a model to begin chatting.

If you picked a first model during setup, it is already downloading, and you can go to Step 4 once it finishes.

  1. Open Models in the sidebar and select the Browse Hugging Face tab.
  2. Type in a popular model name like Gemma3-4B, Qwen3-4B, or gpt-oss-20B and press Enter.
  3. Look for models tagged as GGUF (most common) or MLX (on Apple Silicon Macs).
  4. Click the Download button next to a model variant (Q4KM is usually a good balance of speed and quality).
  5. Follow the progress from the Downloads indicator in the header.

Step 4: Run the Model

Once downloaded, your model will appear in the Local Models library.

  1. Open Models > Local models.
  2. Find your downloaded model and click its Run model button.
  3. The button shows progress while the model prepares. When the model is running, the button changes to Stop model.

Step 5: Start Chatting!

  1. Click on the Chat icon in the sidebar.
  2. Type a message in the text box at the bottom (e.g., "Hello! Can you explain quantum physics in simple terms?").
  3. Press Enter and watch your local AI respond!

Next Steps

Now that you've completed your first conversation, explore more advanced features: