Getting Started

This guide is for first-time users installing InferencePort AI from https://inference.js.org.

Before You Install

  • Confirm your OS: Windows, macOS, or Linux.

  • If you are unsure which download to choose, use Download Guide.

  • If you plan to run local models, make sure Ollama is available.

Install InferencePort AI

  1. Open https://inference.js.org/install.html.

  2. Download the installer/package that matches your OS.

  3. Run the installer (or open the package) and complete setup.

  4. Open InferencePort AI from Start Menu/Applications.

First Launch (Simple Path)

Use this quick path to confirm everything works:

  1. Open Marketplace.

  2. Open Local Chat (Ollama models) and download one model.

  3. Go to Chat, select host and model, then send a test prompt.

  4. Confirm you receive a streaming response.

Optional Setup

You can skip this section and come back later.

  • Sign in (optional): required only for account features and sync.

  • Enable sync (optional): off by default.

  • Add remote hosts (optional): if you use model servers on another machine.

  • Import models (optional): .gguf/Modelfile from Marketplace.

What To Learn Next

Need source build instructions?

If you want to build the app from source code instead of downloading from the website, see Build and Release.