Skip to content

Technology and setup

One download, one password, and you are ready to go.

Snelli is an application for Mac and Windows. No server, no practice network, no IT department required.

What your computer needs

Memory is the one number that matters.

This table answers the most common technical question practices face. Requirements are stated conservatively to ensure smooth performance.

RecommendedMinimum
Memory (RAM)16 GB or more8 GB
Operating systema current macOS or Windowsa current macOS or Windows
Internetonce, for setuponce, for setup

Version 2.1 · As of 8 September 2026

Memory (RAM)

The AI runs on your computer. Speech recognition and the language model never run at the same time, so peak demand is one model. Snelli runs on 8 GB; 16 GB or more gives you headroom, including for long sessions.

Operating system

One application for Mac and Windows. No server and no practice network required.

Internet

For the one-off download of the AI models during setup. Recording, transcribing, writing and exporting all work without it.

Installation

Five things you do exactly once.

Plan roughly 30 minutes before your first session, including the initial download of local AI models.

  1. Download Snelli and install it like any standard application.

  2. On first launch, Snelli downloads the AI models once. You can monitor progress, pause, and resume anytime.

  3. Enter your license key and accept the initial agreements.

  4. Set a master password and securely store your recovery code.

  5. Configure your methods, locations, and working hours. Snelli is then ready to use.

Everyday operation

Updates, connectivity, and working offline.

Updates. You fetch new versions yourself from this website and install them when it suits you. An update never alters your stored data.

Offline mode. You can record, transcribe, generate notes, and export reports without an internet connection. Internet access is needed once, for the initial setup, when Snelli downloads the AI models.

Details on what data is transmitted