How Terraspace Works

Here’s a high-level explanation of how Terraspace works. It’s pretty straightforward.


Terraspace works by building files in the app and config/terraform folders to a .terraspace-cache folder. Here’s a diagram:

Then it merely calls out to terraform within that folder. In fact, you can use Terraspace to build the files first, cd into the .terraspace-cache folder, and run Terraform directly. Example:

terraspace build demo
cd .terraspace-cache/us-west-2/dev/stacks/demo
terraform init
terraform apply

Once you’re in the .terraspace-cache folder, it’s regular terraform at that point.

Terraspace automates it with:

terraspace up demo

HCL ERB Templating: A little extra power

Terraspace supports ERB templating. This minor enhancement adds incredible power.

Terraspace tries to give you the best of both worlds. You get the Terraform HCL declarative power, at the same time, it gives you a little extra power with ERB. Terraspace allows you to reach for this additional power when needed.

If you prefer not to use HCL, then Terraspace also has a Ruby DSL. The DSL builds .json files that Terraform understands just like .tf files.

Backend State Storage

Terraspace can create backend state storage resources like s3 buckets, dynamodb tables, azure storage accounts, and gcs buckets automatically. This added convenience is one less thing you don’t have to worry about. You can focus on code instead.


Terraspace ships with a testing framework. It allows you to test with real resources. The testing works by creating a “test harness”, which is a generated terraspace project. Then terraspace up is ran. It’s what you would normally do manually, but automated.

More tools: