Overview

Morphis separates deployment into two different concerns.

1. Server deploy

Use this mode when you control the machine yourself.

Typical target: Linux VM, VPS, or dedicated server.
Runtime model: build once, then run morphis start as a long-lived process.
Recommended operational layer: systemd, a reverse proxy such as Nginx, and your own logging and monitoring.

Use this mode when you want Morphis to package your server as a Docker image and push it to a managed platform.

Target	Platform model	Docker under the hood	Notes
AWS	Lambda container image	Yes	Builds a Linux amd64 image, adds the Lambda Web Adapter, pushes to ECR, then updates or creates a Lambda function.
Google Cloud	Cloud Run	Yes	Builds a Linux amd64 image, pushes to Artifact Registry, then deploys a Cloud Run service.
Cloudflare	Containers + Workers	Yes	Uses Wrangler and Cloudflare Containers, with a Worker entrypoint and Durable Object routing layer.

morphis deploy --server=api --target=aws

Change the --target value to gcloud or cloudflare for the other managed targets.

Choose a server deploy when you want the simplest operational model and full control over the host.
Choose AWS when you want a container-based Lambda workflow and you are already using the AWS ecosystem.
Choose Google Cloud when Cloud Run is your preferred serverless container platform.
Choose Cloudflare when you want to run behind Workers and optionally attach Cloudflare-native services such as D1.

You can pair each deployment mode with environment-aware files and run migrations before deployment when needed.