This guide shows you how to integrate SambaCloud’s LLMs with the Kilocode VS Code extension. Kilocode is an AI-powered autonomous assistant for planning, building, and fixing code. It is a superset of Roo, Cline, and their own features.

​
Prerequisites

Before you begin, ensure you have:

​
Setup

Follow these steps to connect SambaCloud models in KiloCode:
  1. Open KiloCode
    In VS Code, open the KiloCode panel (usually located in the sidebar).
  2. Add your SambaCloud API key
    1. Click Use your own API key.
    2. In the API Provider dropdown, select SambaNova.
    3. Enter your SambaCloud API key.
    4. Choose a model from the Model dropdown (e.g., Llama-4-Maverick-17B-128E-Instruct).
    5. Select your preferred output verbosity.
    6. Click Let’s go!

​
Usage

  1. Choose an interaction mode
    From the bottom-right panel, select one of the following modes:
    • architect
    • code
    • ask
    • debug
    • orchestrator
    You can customize interaction modes in the Edit section.
  2. Start coding with SambaCloud in KiloCode
    • Provide a task or prompt.
    • Review the assistant’s plan.
    • Approve the actions you want KiloCode to execute.
    Example:
    The model was prompted to build a basic calculator app with a FastAPI backend that performs addition, subtraction, multiplication, and division. The frontend uses HTML/CSS/JavaScript to accept input, select an operation, and display results, communicating with the backend via a fetch API call.     The KiloCode panel displays executed tasks, token usage, and session cost at the top bar.

​
Next steps

Your SambaCloud models now power the KiloCode assistant, delivering inference speeds up to 5x faster than typical GPU providers. If you run into issues, contact your SambaNova support team for assistance.