Installation

NELA is primarily used as a desktop app. Most users should install a release build from the Downloads page, then install models from inside Settings.

The npm flow below is for running NELA from source.

Download from website

You can also install NELA directly from the website Download tab.

1. Open the Download tab.
2. Select your operating system and preferred version.
3. Download the installer and run it.

Install via npm

cd genhat-desktop
npm ci
npx tauri dev

Prerequisites for source runs:

• Node.js 24 or newer
• Rust stable toolchain
• Linux desktop dependencies for Tauri (webkit2gtk, gtk3, asound, and related dev packages)

Install a model (example)

Inside the app:

1. Open Settings from the sidebar.
2. Choose a model category (LLM, Vision, TTS, ASR, Embedding, Classifier, Grader).
3. Download a compatible model.
4. Return to the main view and select that model for your current mode.

Optional for custom model locations in source/dev runs:

export GENHAT_MODEL_PATH=/absolute/path/to/models

Set GENHAT_MODEL_PATH before launching if models are stored outside the default repository models directory.