Desktop

IIMAGINE Desktop

IIMAGINE Desktop

A free, open-source desktop app that runs AI models directly on your computer. No account required, and your data can stay entirely on your machine.

Where to Find It

  • Download page URL path: /downloads
  • Learn more URL path: /open-source
  • Source code: Available on GitHub (linked from both pages)
  • License: Free and open source (MIT)

What You Can Do

  • Download IIMAGINE Desktop for macOS, Windows, or Linux
  • Chat with AI models that run locally on your own computer
  • Browse and download models, with recommendations based on your computer's hardware
  • Switch between models instantly
  • Choose one of three privacy levels for where each request runs
  • Use text-to-speech and voice chat
  • Generate audio from text on a standalone audio page
  • Extend the app with plugins
  • Connect local files and external tools
  • Let other apps on your computer use your local models through a local API

How It Works

Downloading the App

  1. Go to /downloads
  2. You see three download cards — macOS, Windows, and Linux
  3. Each card shows the system requirement (for example, "macOS 12 or later")
  4. Click "Download" on the card for your operating system
  5. The current version and a link to the source on GitHub appear below the cards

Installing

macOS

  1. Open the downloaded .dmg file
  2. Drag IIMAGINE Desktop to your Applications folder
  3. Open it — the app is signed and notarized by Apple, so it opens normally with no security warning

Windows

  1. Run the downloaded installer
  2. If a SmartScreen warning appears, click "More info" then "Run anyway" (the Windows version is not yet code-signed)
  3. Follow the installer to finish

Linux

  1. Make the downloaded .AppImage file executable
  2. Run it — no installation is required

Choosing a Privacy Level

The desktop app offers three levels so you control where each request runs:

  1. Local (maximum privacy) — Models run on your machine. Your data never leaves your device.
  2. Regional Cloud (compliant) — Access GPT, Claude, or Gemini through your organization's approved cloud (Azure, AWS, or GCP) using your own credentials, so data stays within your approved cloud environment.
  3. Third-party Cloud (convenience) — Send requests directly to providers using your own API keys, or through an IIMAGINE subscription.

Chatting With Local Models

  1. Open the app and go to the chat area
  2. Pick a model from the model selector
  3. Type your message and send it
  4. Responses stream back, and your conversation history is kept on your device

Browsing and Downloading Models

  1. Open the models area in the app
  2. Browse the available models — the app recommends models that suit your computer's hardware
  3. Download a model; you can see progress as it downloads
  4. Once downloaded, select it to start chatting

Voice and Audio

  • Turn on voice chat to speak to the AI and hear responses read aloud
  • Use the audio page to type text, pick a voice, generate speech, play it, and download the audio file

Plugins

  1. Open the plugins area in the app
  2. Enable a plugin to turn it on (plugins are off by default)
  3. Some plugins add their own page and settings to the sidebar

Connecting Local Files and Tools

  • Connect a local folder so the AI can work with files on your computer
  • Add external tool connections to extend what the AI can do

Using Your Local Models From Other Apps

The app can expose your local models through a local, OpenAI-compatible connection, so other apps on your computer that support that standard can use your local models.

Requirements

  • Account: None required to download or run the desktop app
  • Operating system: macOS 12 or later, Windows 10 or later, or Ubuntu 20.04+ (or equivalent Linux)
  • Cost: Free

Common Issues

Windows shows a SmartScreen warning when installing → Click "More info" then "Run anyway". The Windows build is not yet code-signed, so this warning is expected.

Not sure which download to pick on Mac → Choose the macOS download. It is built for Apple Silicon (M1 through M4) and requires macOS 12 or later.

Want to run everything with no network access → Use the Local privacy level with a downloaded model. Requests then run entirely on your machine.

Back to Documentation