Skip to main content

Gradient CLI & SDK

Paperspace Gradient CLI & SDK installation and reference guide.


The SDK makes it easy to launch Notebooks, Workflows, and Deployments programmatically using Python.

The SDK is bundled with the Gradient CLI. The Gradient CLI is available on PyPI and works on Windows, MacOS, and Linux.

The CLI requires Python 3.4+ (or Python 2.7). Be sure to use a compatible version of pip (or pip3) depending on your Python version.


We recommend installing and using the CLI within a Python virtual environment. This will minimize conflicts with existing libraries on your computer. We recommend virtualenv. See below for more instructions.

Pip install

pip install -U gradient

The -U option upgrades all specified packages to the newest available version.

You can verify that it is working by running

gradient version

which will output a version like v1.8.6.

Connect your account

You can either stash your API key on your computer or include your API key on each command. The latter can be useful if you are working on several teams at the same time, as each team has its own API key.

Obtaining an API key

Sign in to your Paperspace account and create a new API key. You'll use the API keys you generate here to authenticate your requests.

Set active API key

First, obtain an API Key, and then:


Alternatively, you can set the environment variable PAPERSPACE_API_KEY temporarily to override your configured api key.

gradient workflows run ...

You can reveal your current API key with cat ~/.paperspace/config.json

Using a virtual environment

For Python 3.4+

First, install virtualenv:

pip install virtualenv

Create a new virtual environment:

python3 -m virtualenv <virtual_env_dir_path>

Activate the virtual environment:

source <virtual_env_dir_path>/bin/activate

Virtualenvs can also be run in other ways, e.g., using conda.

Enable autocomplete

Add the following to your .bashrc (or .zshrc) to enable autocomplete anytime you activate your shell. If Gradient was installed in a virtual environment, the following has to be added to the activate script:

eval "$(_GRADIENT_COMPLETE=source gradient)"

Alternatively, you can create an activation script by:

(_GRADIENT_COMPLETE=source gradient) > ~/

and then add . ~/ to your .bashrc, .zshrc or activate script.

For more, see .

Install the latest pre-release version

If you need a pre-release version of the CLI use the following command to install it:

pip install -U --pre gradient