Quickstart

By the end of this page you’ll have H.O.T-Jarvis running on your machine, holding a real conversation through a local model, and remembering it after you quit and reopen the app. It costs nothing.

Prerequisites

• Node.js 20+ and Rust (stable) — the app is Tauri v2 (Rust core, web UI).
• On Linux, the Tauri system dependencies.
• Either Ollama (recommended, local and unlimited) or a free cloud key.

1. Give it something to think with

The assistant needs a model. The private, unlimited, free option is a local one:

install a local model

# install Ollama from https://ollama.com, then:
ollama pull llama3.2

No local model? You can use a free cloud tier instead — see Add a free cloud key. With neither, the app still launches and shows you exactly how to finish setup.

2. Run the app

from the repo root

npm install
npm run tauri dev

The first launch compiles the Rust core, so give it a few minutes. After that, startup is fast.

3. Say hello

Type a message in the composer and send it. The arc-reactor core moves from standby to online once a model is reachable, and the reply’s footer shows which provider and model answered, plus a confidence score.

4. Confirm it remembers

This is the part most “assistants” skip. Tell it something about yourself, then:

1. Quit the app completely.
2. Reopen it.
3. Open the Memory tab.

Your conversation is still there. Memory is a local SQLite database that survives restarts, upgrades, and reboots — and it’s yours to export or wipe whenever you want.

Next

• Author your first skill — watch the assistant write and test a new ability.
• The four hero features — what makes this more than a chat window.