Windows preview  /  v0.3.1

Get TWIN AI.

Install the conversation studio with the Windows setup wizard, or take the portable ZIP.

Windows x64  ·  Setup 18.7 MB  ·  ZIP 18.3 MB

Verify your download.

Setup and its uninstaller use a self-signed TWIN AI Studio preview certificate. Windows may show an unknown-publisher warning.

Installer SHA-256
Verify the download29c3d9bddbef3c76125935eb660f46610d48cdc1235f9ccc3653bad7ce624bf2

Before you start

  • 64-bit Windows and an up-to-date NVIDIA driver for local media models.
  • Plan for 16 GB RAM and 30 GB of free disk space. The development GPU has 8 GB VRAM.
  • An NVIDIA API key for hosted replies. Speech and avatar models install separately.

First launch

  1. With Setup, choose an empty folder and open TWIN AI Studio from Start. For the portable version, extract the complete ZIP into a writable folder.
  2. The first launch prepares the runtime. From the ZIP, start it with Start TWIN AI.cmd.
  3. Set NVIDIA_API_KEY in the local .env file, restart, and open Setup to choose your media components.

Use Windows Installed apps or the installed uninstall.exe to remove the studio. Personal profiles and model files stay in place. Tested hardware, known limits, and third-party model terms are listed in the included documentation.

Questions before installing

Does this AI twin run locally?

The Windows app runs the speech and avatar models on your PC. It sends the question and relevant profile text to NVIDIA for the written reply. An internet connection and NVIDIA API key are required for that step.

Can I use it with just my CV and project notes?

Yes. Create a profile, add your background, and choose text-only replies. Type a question or use selected text to try the studio before recording a voice or avatar.

How do I send my avatar and voice to a call?

Prepare an avatar session, then select Unity Video Capture as the camera and CABLE Output as the microphone in the receiving app. The device guide shows how the inputs and outputs connect.