Skip to content

Quick start

This guide gets you from download to a working AI-assisted terminal in about five minutes.

Download a build for your platform, or install from a package manager. Full details live in the installation guide:

  • Linux.deb, .rpm, AppImage, or the AUR package. See Linux.
  • macOS.dmg for Apple Silicon or Intel. See macOS.
  • WindowsNSIS or MSI installer, plus first-class WSL. See Windows & WSL.

Launch Nexis and open a folder — File → Open Folder, or drag a directory onto the window. The file explorer, source control, and terminal all scope to that workspace. A terminal tab opens in the project root automatically.

Drop a NEXIS.md file in the project root to give the AI persistent project memory — conventions, architecture notes, or anything you want it to remember across sessions.

Open Settings → AI:

  • Bring your own key — pick a provider (OpenAI, Anthropic, Google, Groq, and more) and paste your API key. Keys are stored in the OS keychain — never on disk.
  • Run offline — point Nexis at a local LM Studio, MLX, or Ollama endpoint. No key required, nothing leaves your machine.

Toggle the AI panel any time with the AI panel shortcut (see keyboard shortcuts).

  • Type a command; select its output and click Ask Nexis or Explain to send it to the AI chat.
  • Open a file, start typing, and accept an inline autocomplete suggestion.
  • Ask the AI to make a change — it proposes edits as per-hunk diffs you approve or reject individually.
  • Stage your changes in Source Control and let the AI draft a Conventional Commit message from the diff.