Get started

From nothing to your first conversation.

Pick Aivyx PA (a personal assistant), aivyx-coder (a coding agent in your terminal), or both. Each takes a few minutes.

1. A model to talk to

Both run on a model on your own computer. If you don't have one yet, Ollama is the quickest start (from ollama.com; on Linux:)

curl -fsSL https://ollama.com/install.sh | sh
ollama pull qwen3:8b

Lemonade, llama.cpp's llama-server, Jan and Docker Model Runner work too; setup finds whichever is running. Aivyx PA can also use an Anthropic or OpenAI account under your own key instead; aivyx-coder is local only. A bigger model plans and uses tools better: see models and routing.

2a. Aivyx PA

On Linux or macOS, in a terminal:

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/Aivyx-Agent/aivyx-pa/releases/latest/download/aivyx-cli-installer.sh | sh
aivyx-pa init    # finds your model, sets it up, checks it works

aivyx-pa init finds your model, asks a few questions about you and what you want help with, and sets a passphrase that keeps its memory encrypted. Then aivyx-pa chats in the terminal, andaivyx-pa studio gives you the link to the Studio in your browser. If your shell says command not found, open a new terminal first.

Next: Getting started in the manual.

2b. aivyx-coder

On Linux (x86_64) or an Apple Silicon Mac, in a terminal:

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/Aivyx-Agent/aivyx-coder/releases/latest/download/aivyx-coder-installer.sh | sh
cd your-project && aivyx-coder    # finds your model, sets it up, opens the TUI

Then, in your project's folder, aivyx-coder: the first run finds your model server, lets you pick a model, checks that it answers and writes your config. Describe a change; it reads your project, edits, runs your tests and asks before each step.

Next: Install and first run in the manual.

3. Using both

They're separate programs with their own setup, and both can use the same model server. Aivyx PA can also hand coding jobs to aivyx-coder (the recipe ↗).