Private beta · Platform control plane

Docker Compose to a live URL — in your own AWS account.

plactl is a platform control plane that keeps up with the dramatic increase in developer — and AI-agent — productivity. Point it at the compose file your team already runs locally; get production infrastructure with the guardrails a platform team would have built.

Terminal demo of plactl. Three commands run in sequence: brew install plactl, plactl login, and plactl connect dev. The output shows AWS credentials being set, the kubeconfig updated to a plactl-dev context, port forwards opened to postgres on localhost 5432 and redis on localhost 6379, and a DATABASE_URL secret injected into the developer's environment.

Built for both

Two audiences. One product.

For the buyer

  • Hiring a platform team: $750K–$1.5M / year (3–5 FTEs, 6-month hire, 6-month ramp). plactl is a fraction of that, in days.
  • Platform setup: 30–45 minutes. First deploy: minutes after that.
  • AWS account structure, GitOps wiring, Kubernetes, observability — provisioned automatically.

For the developer

  • plactl connect dev — credentials, kubeconfig, port forwards, secrets. One command.
  • No AWS console. No IAM configuration. No asking for the database password.
  • Push the compose file you already run locally, get a live URL. Build and deploy status lands on the PR; logs deep-link into your own AWS console.

How it works

Six steps. About thirty minutes.

  1. Install the GitHub App. Point it at your org.

  2. Create a platform. Define your environments (dev, staging, prod).

  3. plactl provisions AWS accounts, Kubernetes clusters, CodeBuild runners, and GitOps wiring — automatically, in about 30 minutes.

  4. Developers run plactl connect dev. They get credentials, a kubeconfig, and live port forwards to every service in dev.

  5. They push code. plactl reads the compose file, builds and deploys the services, and posts status on the PR — ending in a live URL.

  6. They promote to production with plactl promote staging prod.

No Terraform knowledge required. No platform engineer required.

A way to put it

Your AWS. Their developer experience.