Skip to main content

Prerequisites

Before you install and run azuma nori, make sure you have everything below ready. The whole analysis runs locally on your workstation — your source code never leaves your machine.

1. A supported workstation​

azuma nori ships as a desktop application for:

  • Windows 10 / 11 (64-bit)
  • macOS — Apple Silicon (arm64) or Intel (x64)
  • Linux — Ubuntu 24.04+ (the signed .deb package is the supported install; an AppImage is also provided)

2. An azuma account​

Sign-in uses your azuma doa identity (OpenID Connect). You don't need one in advance — if you don't have an account yet, you can create one (and your own workspace) right from the desktop app's sign-in screen. See Create an account.

3. An AI model​

nori performs its analysis through a Large Language Model that you connect. You can use either of two connection types:

  • Direct Cloud API — a provider API key for OpenAI, Azure OpenAI, Claude, or Gemini (which uses Google's native Gemini API for direct connectivity, supporting Gemini 3 thinking models with native token tracking).
  • CLI / Local Executable — a locally installed agent CLI (Claude, Gemini, Cursor, or Codex) that nori drives as a local process. Use this to keep code evaluation on your own infrastructure (for example, private-model or air-gapped setups).
Keys stay local

All configured models are saved locally on your machine and are never sent to the cloud backend.

4. The code you want to audit, and a data directory​

  • The source code repository you want to analyze, available as a local folder (your Root Directory).
  • A separate, writable Data Directory where nori stores its configuration, artifacts, and analysis results (the local .nori workspace). Keep this distinct from your source tree.

Once these are in place, continue to Installation.