- Linux (Debian/Ubuntu fully supported; other distros can diagnose but not auto-apply fixes)
- Python 3.8 or newer
- Ollama — installed automatically during setup if not present
Check your Python version:
python3 --versionIf it is below 3.8 or missing entirely, see Installing Python.
curl -fsSL https://raw.githubusercontent.com/JayRaj21/DaemonIQ/master/install.sh | bashThe script checks for Python, downloads the DaemonIQ files, creates the daemoniq command in ~/.local/bin, and adds it to your PATH. The only prompt it asks is whether you want the demon to start automatically at login.
When it finishes, open a new terminal before running daemoniq. This is necessary for the PATH change to take effect.
Download install.sh, daemoniq-imp.py, and daemoniq-sovereign.py into the same directory, then run:
bash install.shThe installer looks for the script files in the same directory, so no internet connection is needed.
Running daemoniq for the first time starts a short setup wizard. It asks two questions.
Question 1 — which model tier?
1) Imp Runs on most machines with 4GB+ RAM.
Uses Llama 3 via Ollama (~4GB download).
2) Demon Best local quality. Needs 9GB+ RAM.
Uses Qwen2.5 via Ollama.
If you are unsure, choose Imp. It works on most machines and the quality is good enough for the majority of troubleshooting tasks. You can switch later with daemoniq setup.
Question 2 — RAM size (Demon only)
Qwen2.5 comes in several sizes. The wizard maps your RAM to the right one:
| Your RAM | Model used | Download size |
|---|---|---|
| 8–12 GB | qwen2.5:7b | ~5 GB |
| 12–24 GB | qwen2.5:14b | ~9 GB (recommended) |
| 24 GB+ | qwen2.5:32b | ~20 GB |
| Not sure | qwen2.5:7b | ~5 GB |
Both Imp and Demon use Ollama to run the model locally. If Ollama is not installed, the setup wizard will offer to install it for you. To install it manually:
curl -fsSL https://ollama.com/install.sh | shTo pull a model manually instead of through the wizard:
ollama pull llama3 # Imp
ollama pull qwen2.5:14b # Demon — adjust tag to match your RAMOllama resumes interrupted downloads, so if a large model download fails partway through, simply run the pull command again.
To switch between Imp and Sovereign, change your distro selection, or swap models:
daemoniq setupDebian / Ubuntu / Mint / Pop!_OS
sudo apt update && sudo apt install python3Fedora
sudo dnf install python3Arch / Manjaro
sudo pacman -S pythonAfter installing, close and reopen your terminal before running the installer again.
command not found: daemoniq after a successful install
The PATH change written by the installer only takes effect in new terminals. Open a new terminal window, or reload your config manually:
source ~/.bashrc # bash
source ~/.zshrc # zshdaemoniq starts but immediately exits with a Python error
Check the demon log:
daemoniq logsOllama is not running
ollama serveTo start it automatically on login:
systemctl --user enable --now ollamaSlow or interrupted model download
Run the pull command again — Ollama resumes from where it stopped:
ollama pull llama3
ollama pull qwen2.5:14bSomething else
daemoniq logsThe log file at /tmp/daemoniq-demon.log records everything the demon does and is usually the fastest way to diagnose unexpected behaviour.
daemoniq uninstallThis stops the demon, removes all installed files, deletes the daemoniq command, disables the systemd service if one was created, and removes the PATH entry from your shell config. Open a new terminal afterwards to confirm the command is gone.
All files are installed to your home directory. Nothing is written system-wide.
| Path | Contents |
|---|---|
~/.daemoniq-demon/ |
Scripts, config, backups, hardware snapshot |
~/.daemoniq-demon/config.json |
Your setup choices |
~/.daemoniq-demon/hardware_snapshot.json |
Hardware scan from last demon start |
~/.daemoniq-demon/backups/ |
Pre-patch backups (up to 5 per variant) |
~/.local/bin/daemoniq |
The shell command |
/tmp/daemoniq-demon.log |
Demon log (cleared on reboot) |
/tmp/daemoniq-demon.sock |
Unix socket (cleared on reboot) |