Getting started

You need a Facets control plane, a cloud account and the Praxis CLI. Your AI assistant can then do most of the setup, or you can do each step yourself.

Your first running service

Six steps. Once your cloud account is connected and modules are loaded, plan on 15 to 30 minutes, most of it the first environment launch.

1. Get a control plane: Facets hosted, or self-hosted in your own cloud account with the Facets team.
2. Install the Praxis CLI and log in. praxis login also installs raptor, the Facets command line, and adds the Facets skills to your AI assistant. On Linux, download the binary instead (Install Praxis CLI).

   brew install --cask Facets-cloud/tap/praxis
   praxis login --url <your-control-plane-url>

3. Connect a cloud account: AWS, GCP or Azure, or a Kubernetes cluster (Integrating Cloud Accounts). A Git account (GitHub, GitLab or Bitbucket) for each project's blueprint is optional at first.
4. Load modules and create a project type, which decides the modules a project can use. Loading modules from Facets' catalog needs git and Terraform on your machine.
5. Create a project, add a service and launch an environment.
6. Release. The environment shows Running, with the service's status, logs and metrics.

Your AI assistant can run these steps for you (Deploy with an AI agent), or you can run them by hand (Deploy your first application).

Trying it before you connect your cloud

You can do everything up to launching an environment before you connect a cloud. Sign up for Facets hosted with a work email and no credit card, for a 14-day free trial, then load modules and design a blueprint. To look around without signing up, click through the product demos or book a demo.

Adopting one piece at a time, and how long it takes

You can bring in one service or one environment at a time. What you haven't moved keeps running where it is.

• A service: point it at the image your CI builds.
• Running resources: import them without changing them.
• Your Terraform modules: wrap them in a Facets module.
• Resources you keep outside: reference them in the blueprint.

How long it takes depends on how many resources you import, and how many need a module you don't have yet. For example:

• GGX migrated to GCP in two weeks instead of the expected 2-3 months, a 75% time reduction (case study)
• Spectrum moved its Azure event processing to Google Cloud in two weeks with no customer impact; the AWS backend followed in 90 days (case study)