Need help with susi-assistant?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

131 Stars 0 Forks Apache License 2.0 1.1K Commits 0 Opened issues


SUSI Smart Assistant

Services available


Need anything else?

Contributors list

Installation of SUSI.AI Personal Assistant

This repository provides installation packages for the SUSI.AI Personal Assistant, both for desktop installation as well as images for the SUSI.AI Smart Speaker based on Raspberry Pi.


If you want to get SUSI.AI Personal Assistant installed quickly without reading all the small print, here are two avenues you can select: Either use our pre-built images, or install using our scripts.

Installation via install script

Run the following command (assuming you have

curl --proto '=https' --tlsv1.2 -sSf | bash
If you haven't
installed but
, you can use the following line instead:
wget -O - | bash

Note that this will use

and you will get asked for the password.

This will be default install SUSI.AI into

, adds desktop entries as well as systemd unit files for starting and stopping the various services.

Installation via pre-built images

WARNING we strongly suggest using the method above. If you really want to use some different method, please see the various install scripts.

Step 1: Download the SUSI.AI Smart Assistant package

Get the installer package SUSIAI-release-YYYYMMDD.S.tar.gz (or SUSIAI-build-YYYYMMDD.S.tar.gz) from the Github SUSI Installer release page.

Step 2: Unpack the package to your preferred location

The package unpacks into a directory

which can be placed anywhere in your system. We generally recommend placing it into
but feel free to move it somewhere else.

Step 3: Install necessary dependencies

The SUSI.AI Smart Assistant depends on a lot of additional software that can be installed using the included script

bash --system-install --with-deepspeech
For other options, please see below for details.


is not an option, the following method can be used
bash --sudo-cmd ""
This method is much less tested, because the required Python modules will be installed into your home directory (
). On the other hand, no
permissions are needed, and everything can be done as local user.

Optional step 1: Linking start/stop scripts

There are several script to start, stop, and configure SUSI.AI available in

, but this directory is usually not in your PATH environment variable.

If you don't want to call the scripts always with full path, you can either add

to your PATH
PATH=.../SUSI.AI/bin:$PATH ; export PATH
or link the scripts to one of the directories already in your PATH
.../SUSI.AI/bin/susi-config install links 
 is the place where the links are created.

So for example

.../SUSI.AI/bin/susi-config install links ~/bin
would make the SUSI.AI scripts available in

Optional step 2: Adding entries to the DE menus

You can add menu entries to your desktop environment by calling

.../SUSI.AI/bin/susi-config install desktop user
Having done this, the following items should be available in your desktop environment menu:
  • SUSI Server - starts the SUSI Server, a necessary component
  • SUSI.AI Personal Assistant - starts the privacy assistant in the background
  • SUSI.AI Personal Assistant - Application Window - an application that allows interaction via an GUI
  • SUSI.AI Personal Assistant - Configuration - configuration of various parameters of SUSI.AI

More programs and services will be added over time.

Optional step 3: Adding Systemd integration

Systemd integration can be achieved by installing several .service files into the respective locations using

.../SUSI.AI/bin/susi-config install systemd user

After installation of the systemd service files, the SUSI.AI Personal Assistant can be started once by typing:

systemctl --user start ss-susi-server
systemctl --user start ss-susi-linux
If you want to enable the assistant permantenly on your desktop, use
systemctl --user enable ss-susi-server
systemctl --user enable ss-susi-linux

Available options for the installation scripts

For more controlled installation with lots of options to be configured, see below for the detailed explanation of the installation scripts.

Starting and stopping SUSI.AI

To start the SUSI.AI Smart Assistant, first you need to start the SUSI.AI Server, by either calling

$HOME/ start
or, if systemd integration was installed, by calling
systemd --user start ss-susi-server

After this, you are ready to start SUSI.AI Smart Assistant by either calling

$HOME/ start
or, if systemd integration was installed, by calling
systemd --user start ss-susi-linux

You can stop the two service in the similar way:

$HOME/ stop
$HOME/ stop
or, if systemd integration was installed, by calling
systemd --user stop ss-susi-linux
systemd --user stop ss-susi-server


See details in

Installation of requirements

The provided script

checks that the above mentioned programs are available, checks the version number of the available
binary, updating it if necessary using
itself, and then uses
to install the missing requirements.

The script uses

to obtain
rights to install the necessary Python libraries.

Possible options:

  --trust-pip      Don't do version checks on pip3, trust it to be new enough
  --branch BRANCH  If no local checkouts are available, use the git remotes
                   with branch BRANCH to get requirement files (default: development)
  --raspi          Do additional installation tasks for the SUSI.AI Smart Speaker
  --sudo-cmd CMD   Use CMD instead of the default sudo
  --system-install Try installing necessary programs, only supported for some distributions
  --sys-installer ARG   Select a system installer if not automatically detected, one of "apt" or "dnf"
  --with-deepspeech Install DeepSpeech and en-US model data
  --no-install-node Don't install node and npm from NodeSource
                   If Node and NPM are available in sufficiently new versions,
                   no update/install will be done anyway
  --no-clean       Don't remove temp directory and don't use --no-cache-dir with pip3
  --quiet          Silence pip on installation

Installation of SUSI.AI

The installaction script
carries out the actual installation on the target system. It checks that the above set of programs is availabe, clones the necessary git repositories from github, and installs Systemd unit files to allow starting/enabling the respective programs. On the Raspberry based SUSI.AI Smart Speaker, several further tasks are carried out to set up audio playback, sound device setup etc.

Installation modes

The installation can be carried out in either

moder or
  • user mode
    : All files are installed into the directory
    in the home directory of the current user (the location can be configured). Systemd unit files into the current user's systemd configuration directory. Utility programs for SUSI.AI are installed into
  • system mode
    : All files are installed into the directory
    (again, the location can be configured), and systemd unit files are installed in the system-wide systemd configuration directories. Utility programs for SUSI.AI are installed into


system mode
a new dedicated user for the SUSI.AI server process is necessary, and will be created. By default it will be named

Command line options for

The behaviour of the installation script can be changed with the following command line options:

  --system         install system-wide
  --prefix    (only with --system) install into /lib/SUSI.AI
  --destdir   (only without --system) install into 
                   defaults to $HOME/SUSI.AI
  --susi-server-user  (only with --system)
                   user under which the susi server is run, default: _susiserver
  --dev            use development branch
  --with-coral     install support libraries for the Coral device (Raspberry)

Configuration file

During installation a configuration file is generated in

, which normally is
. This file is read from various parts of the SUSI.AI system. Changes to this file are preferrably made by calling
which is installed into

SUSI.AI Smart Speaker

During installation of the SUSI.AI Smart Speaker, further tasks are performed, and the behaviour cannot be changed using command line options.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.