# CDKTF — Define Terraform Infrastructure with TypeScript, Python, Go or Java

> CDK for Terraform (CDKTF) lets you define cloud infrastructure using familiar programming languages instead of HCL, then synthesizes standard Terraform JSON for plan and apply.

## Install

Save in your project root:

# CDKTF — Define Terraform Infrastructure with TypeScript, Python, Go or Java

## Quick Use
```bash
# Install the CLI
npm install -g cdktf-cli

# Initialize a new TypeScript project
cdktf init --template=typescript --providers=aws

# Write infrastructure in TypeScript, then deploy
cdktf synth   # generates Terraform JSON in cdktf.out/
cdktf deploy  # runs terraform apply under the hood
```

## Introduction
CDK for Terraform bridges the gap between Terraform's provider ecosystem and general-purpose programming languages. Instead of writing HCL, you define resources with TypeScript, Python, Java, C#, or Go classes, gaining access to loops, conditionals, type checking, and IDE support while Terraform handles the actual provisioning.

## What CDKTF Does
- Generates Terraform-compatible JSON configuration from imperative code
- Provides type-safe constructs for every Terraform provider and module
- Supports all existing Terraform providers without modification
- Enables code reuse via functions, classes, and package managers (npm, PyPI, Maven)
- Works with Terraform Cloud and Terraform Enterprise for remote state and runs

## Architecture Overview
CDKTF uses the AWS CDK's construct programming model. You write stacks of constructs in your chosen language, and `cdktf synth` walks the construct tree to emit Terraform JSON into `cdktf.out/`. From there, `cdktf deploy` invokes the Terraform CLI for planning and applying. Provider bindings are auto-generated from Terraform provider schemas via `cdktf get`, producing language-native classes with full type information.

## Self-Hosting & Configuration
- Install the CLI globally via npm; language-specific project templates handle SDK installation
- `cdktf.json` at the repo root lists required providers, modules, and Terraform backend config
- `cdktf get` generates typed provider bindings into a local `.gen` directory
- Remote backends (S3, GCS, Terraform Cloud) are configured in the stack definition
- CI/CD pipelines run `cdktf synth` then standard `terraform plan/apply` for review workflows

## Key Features
- Full IDE support with autocompletion, refactoring, and go-to-definition for infrastructure code
- Access to the entire Terraform provider ecosystem (AWS, GCP, Azure, Kubernetes, and hundreds more)
- Construct-level abstraction allows publishing reusable infrastructure libraries
- Unit testing with standard language test frameworks (Jest, pytest, JUnit)
- Incremental adoption: import existing Terraform state with `cdktf import`

## Comparison with Similar Tools
- **Terraform (HCL)** — Declarative DSL with a mature ecosystem; CDKTF adds imperative flexibility on top
- **Pulumi** — Similar concept but uses its own state backend and provider model; CDKTF reuses Terraform providers directly
- **AWS CDK** — Imperative IaC for AWS CloudFormation; CDKTF applies the same pattern to any Terraform provider
- **OpenTofu** — Community fork of Terraform; CDKTF can target OpenTofu as well
- **Crossplane** — Kubernetes-native infrastructure management via CRDs; different operational model

## FAQ
**Q: Can I mix CDKTF and HCL in the same project?**
A: Not directly in the same stack, but you can reference HCL modules from CDKTF stacks using `TerraformHclModule`.

**Q: Does CDKTF support Terraform state locking and remote backends?**
A: Yes. CDKTF stacks configure backends the same way Terraform does, including S3, GCS, Azure Blob, and Terraform Cloud.

**Q: How do I test my CDKTF infrastructure code?**
A: CDKTF provides a `Testing` utility that lets you snapshot-test synthesized output or assert on specific resource properties using your language's test framework.

**Q: Is CDKTF production-ready?**
A: HashiCorp considers CDKTF generally available since version 0.15+. It is used in production by teams that prefer imperative languages over HCL.

## Sources
- https://github.com/hashicorp/terraform-cdk
- https://developer.hashicorp.com/terraform/cdktf

---
Source: https://tokrepo.com/en/workflows/asset-67452a98
Author: AI Open Source