» Tokens

Tokens represent values that are unknown until Terraform applies your configuration. For example, names of cloud resources are only assigned upon creation.

Some attributes specified using CDK for Terraform (CDKTF) may not directly map to the values required for Terraform configurations. You can use Tokens to cast these attributes to the correct Terraform language syntax.

» Use Tokens

You may need to use Tokens for:

  • Module outputs for boolean, string, lists, maps, and other complex types.
  • Resource attributes (such as id).
  • Terraform outputs based on resource attributes.

» Example

An EKS module requires a list of subnet ids in order to create a cluster. The VPC module outputs a list of subnets.

To pass the subnet id list to the EKS module, you can use publicSubnetsOutput to retrieve the list from the VPC. However, the subnets attribute requires a list of strings. Use Token.asList(vpc.publicSubnetsOutput) to cast the interpolated module output as a list of strings.

const logRetention = new TerraformVariable(this, "logRetentionInDays", {
  type: "number",

const vpc = new Vpc(this, vpcName, {
  name: vpcName,
  publicSubnets: ["", ""],

new Eks(this, "EksModule", {
  clusterName: "my-kubernetes-cluster",
  subnets: Token.asList(vpc.publicSubnetsOutput),
  clusterLogRetentionInDays: logRetention.numberValue,

Initially, CDKTF will resolve Token.asList(vpc.publicSubnetsOutput) to ["#{TOKEN[TOKEN.9]}"] and logRetention.numberValue to a big negative number like -123828381238238. Later in synthesis, CDKTF will resolve the token to ${module.<module id>.public_subnets} and ${var.logRetentionInDays}.

  "module": {
    "helloterraEksModule5DDB67AE": {
      "cluster_name": "my-kubernetes-cluster",
      "subnets": "${module.helloterraMyVpc62D94C17.public_subnets}"

Refer to the AWS CDK documentation for more detailed information about tokens.