Warning: This document is for an old version of Rasa X.

One-Line Deploy Script

The simplest and quickest way to deploy Rasa X on a server or cluster is to use our one-line deploy script. Simply run the command below in a compatible (see Hardware & OS Requirements) deployment environment (e.g. a Google Cloud VM).

curl -s get-rasa-x.rasa.com | sudo bash

If you get stuck, email support@rasa.com with subject “deployment help” and we’ll jump on a call with you! We’re working to make deployment easier and would love your feedback.

The script will perform the following steps:

  1. Install an embedded Kubernetes cluster. If the Kubernetes command-line interface kubectl is already configured, it will use the configured cluster. Please see Existing Cluster Installation for more details if you are interested in using the script with an existing Kubernetes cluster.

  2. Install the Helm command-line interface.

  3. Deploy Rasa X using the Rasa X Helm Chart.


Before installing using the script you’ll need a deployment environment that can run Rasa X. Please see the Hardware & OS Requirements for the requirements.

Once you are ready, run the install command:

curl -s get-rasa-x.rasa.com | sudo bash


To update an existing deployment to the latest version of Rasa X, simply run the installation command from above again.

To update or downgrade to a specific version of Rasa X, use the RASA_X_VERSION environment variable described in the Customizing the Script section.

Accessing Secrets

To access auto-generated passwords later on, head over to the cluster deployment docs on Accessing Secrets. Both the release_name and namespace is rasa.

Hardware & OS Requirements

The installation script is officially supported on the following operating systems:

  • Ubuntu 16.04 / 18.04

  • Debian 9 / 10

  • Red Hat 7 / 8

  • CentOS 7 / 8

For any other Linux operating systems, please follow the Docker-Compose Manual Install. Windows and MacOS are not currently supported.

Embedded Cluster Installation

In case you don’t have an existing Kubernetes Cluster, the installation script will set up an embedded Kubernetes cluster on your machine. To run Rasa X on your machine, it should have the following hardware specs:


  • Minimum: 2 vCPUs

  • Recommended: 2-6 vCPUs


  • Minimum: 4 GB RAM

  • Recommended: 8 GB RAM

Disk Space

  • Recommended: 50 GB disk space available

Existing Cluster Installation

You can also use this installation method with an existing Kubernetes cluster. The existing cluster has to configured with the Kubernetes command-line-interface kubectl. To see if kubectl is installed on your machine and if it is correctly connected to a cluster run the following command:

# Since the installation script runs as `root`,
# you have to check the `kubectl` context for the `root` user
sudo kubectl version --short

# The output should be similar to this
# Client Version: v1.17.3+k3s1
# Server Version: v1.17.3+k3s1

If you get an error kubectl: command not found or there is no output for the Server Version either connect to a Kubernetes cluster or use the Embedded Cluster Installation.

Once you are correctly connected to the cluster, simply run the installation command:

curl -s get-rasa-x.rasa.com | sudo bash

Customizing the Script

Applying the Settings

You can customize the installation script using the following environment variables. E.g. to configure the Rasa X user password, add the following:

export INITIAL_USER_PASSWORD="my-safe-password"
# -E flag applies environment variables from the current user for `sudo`
curl -s get-rasa-x.rasa.com | sudo -E bash


The deploy script provides a quick deployment to get started with and hence provides only essential customization options. If you decide that you want to to have more control over the deployment (e.g. more configuration options), you can always switch use the Rasa X Helm Chart directly with your currently used cluster.

Available Settings

  • INITIAL_USER_PASSWORD (generated by default): The password to log into Rasa X.

  • ACTION_SERVER_IMAGE (default: an empty placeholder image): The Rasa action server image which runs your custom actions server.

  • ACTION_SERVER_TAG (default: ""): The tag of your action server image.

  • ADDITIONAL_CHANNEL_CREDENTIALS: Messaging and voice channels which can be used to send messages to Rasa Open Source. Examples:

    • REST:

    • Facebook:

      export ADDITIONAL_CHANNEL_CREDENTIALS='facebook.verify="rasa-bit",facebook.secret="secret",facebook.page-acess-token="token"'
  • RASA_X_VERSION (default: 0.26.4): Version of Rasa X to install.

  • RASA_VERSION (default: currently latest version): Version of Rasa Open Source to install.

  • DEBUG_MODE (default: false): Runs the deployment in debug mode which provides more information to debug the deployment in case of problems.

  • ENABLE_DUCKLING (default: False): Define if the deployment should include Duckling.