Install Jaseci

Jaseci is tested and supported on the following systems:

  • python 3.10 or higher
  • Ubuntu 18.04 or later
  • Windows WSL2
  • macOS* (see note below for compatibility with Apple silicon)


Jaseci requires the following dependencies:

  • python3.10-dev
  • g++
  • build-essential
  • pkg-config
  • cmake

To install in Debian (or WSL2),

apt-get install python3.10-dev python3-pip git g++ build-essential pkg-config cmake


For macOS, install the above dependencies using one of macOS package manager such as Homebrew and MacPorts.

Install via pip

Install Jaseci with Python's pip package manager

# Upgrade pip to the latest
pip install --upgrade pip

# Core Jaseci
pip install jaseci

# Jaseci server
pip install jaseci-serv

To check for succesfull installation, execute in terminal

jsctl info

The following output should show.

  "Version": "",
  "Creator": "Jason Mars and friends",
  "URL": ""

Additional Pacakges for AI Modules

jaseci and jaseci-serv packages provide the core of the Jaseci framework. To build jaseci program with AI modules, you need to install additional pacakges.

# Jaseci AI modules for Natural Language Processing (NLP)
pip install jac_nlp[all]

# Jaseci AI modules for Speech
pip install jac_speech[all]

# Jaseci AI modules for Computer Vision
pip install jac_vision[all]

# Other Jaseci AI modules
pip install jac_misc[all]

You do not need to install all of the above packages or even everything in a specific jac_* package. You can cherry-pick specific modules to install based on what is needed for your application. Details on which module is included in each package and how to install selectively can be found here


For macOS, there is currently a known compatibility issue between tensorflow-text and Apple custom ARM-based silicon (M1, M2, etc.). If you are on a Mac machine with an Apple chip, you can still use jaseci and jaseci-serv and majority of the AI modules come with Jaseci, with the exception of those depending on tensorflow-text, which includes use_enc and use_qa in the jac-nlp package.

Alternatively, you can build tensorflow-text from source following solutions provided by the community.

Upgrade Versions

To upgrade installed version of Jaseci core packages to the latest version from Pypi

pip install --upgrade jaseci jaseci_serv

Similarly, to upgrade the jaseci AI kit packages from Pypi

pip install --upgrade jac_nlp[all]
pip install --upgrade jac_speech[all]
pip install --upgrade jac_vision[all]
pip install --upgrade jac_misc[all]

To install specific version of Jaseci

pip install jaseci==

Build from Source

If you wish to use the development version of Jaseci, you can download the source code from Github and build from source.

git clone
cd jaseci/jaseci/ && source
cd jaseci/jaseci_serv && source
cd jaseci/jaseci_ai_kit && source all

For Contributors

If you'd like to make contribution to Jaseci Open Source, you should build from source. In addition, you should set up the following in your development environment to follow the Jaseci Open Source Code Standards.

# Install black
pip install black
# Install pre-commit
pip install pre-commit
pre-commit install

You'll need to add --max-line-length=88 --extend-ignore=E203 arguments to flake8 for linting. We recommend setting it up in your preferred code editor or IDE, e.g. VSCode.

The Jaseci Open Source Contribution Guidelines can be found here.

Running a Jaseci Container

The Jaseci Docker images are built with the dependencies and the Jaseci package installed. A Docker container runs in a virtual environment.

docker pull jaseci/jaseci:latest # Download the latest Jaseci image
docker run -it jaseci/jaseci:latest /bin/bash # Start the container and launch an interactive terminal inside it

This will open up a terminal inside the running container.

We also provide several other Docker images. These images include the core jaseci installation as well as one of the jaseci AI kit package. For example, jaseci/jac-nlp have all modules in jac_nlp installed.

docker pull jaseci/jac_nlp:latest
docker run -it jaseci/jac_nlp:latest /bin/bash