Skip to content

Deploying on AWS

This guide provides step-by-step instructions for deploying Refact on AWS.


  • An AWS account

Log into AWS

  1. Navigate to the AWS Console.
  2. Log in using your credentials.
  3. Locate and select the EC2 Service to access the EC2 Dashboard.

AWS Console

Create a Key Pair

  1. In the EC2 Dashboard, find the Key Pairs option in the sidebar under Network & Security.
  2. Click on Create Key Pair.
  3. Name your key pair (e.g., refact_id_rsa) and create it.
  4. Download the key pair file (.pem file) to your computer.

Key Pair

Launch an EC2 Instance

  1. Return to the Dashboard and click Launch Instance.
  2. For the AMI selection, choose an Ubuntu server image that includes NVIDIA drivers. Note: Ignore the AMI options that come with PyTorch and TensorFlow pre-installed. We’ll use Docker to manage all necessary packages. Launch Instance
  3. Click Compare instance types. Use the filter to select a GPU machine type according to your requirements. Compare
  4. Proceed to select your previously created key pair during instance setup.
  5. Under Network settings and Storage, you can stick with the default configurations or adjust according to your needs.
  6. Click Launch Instance to deploy your Ubuntu server.


Access Your EC2 Instance

  1. After the instance is up and running, navigate to Instances in the sidebar to find your new machine. AWS Machine
  2. Locate the Public IPv4 DNS address for your instance (e.g.,
  3. Open a terminal on your computer and establish an SSH connection with the command:
ssh -i "/path/to/your/refact_id_rsa.pem"

Replace "/path/to/your/refact_id_rsa.pem" with the actual path to your downloaded key pair file.

Deploy Refact

Once connected via SSH, follow the instructions provided by the CLI to complete the setup and deployment of Refact.

Refact CLI

Continue to the Refact Enterprise page to learn how to use Refact for teams.