## The short version

To deploy a Node application, make sure it has a production start path and listens on the PORT environment variable at 0.0.0.0. Ample detects common Node projects, installs dependencies, runs the build when present, selects a Node 22 runtime for a new server, and waits for the app to pass health checks.

## Before you start

This guide is for Node, Express, Fastify, and other JavaScript server projects that need a public URL.

## How to do it

1.  
   ### Define a production start command
   The process should start the actual web server, not a file watcher or development server. Put the command in package.json or provide it explicitly.
   
2.  
   ### Bind to the platform port
   Read process.env.PORT, fall back only for local use, and listen on 0.0.0.0 so the platform can reach the process.
   
   ```
   server.listen(Number(process.env.PORT || 3000), '0.0.0.0')
   ```

3.  
   ### Run the plan
   The plan can detect linked databases, required secrets, and whether the repo contains multiple services.
   
   ```
   ample plan --write
   ```

4.  
   ### Deploy and inspect the result
   Apply the plan. If activation fails, use runtime logs to inspect the process output rather than rerunning the same deployment.
   
   ```
   ample up
   ```

## What this supports

### Good fit
- Node 22 applications
- Express and Fastify APIs
- npm, pnpm, yarn, and bun dependency workflows
- Linked managed databases and persistent volumes

### Know the limits
- Development-only start commands are not suitable for production
- A Node app must use a Node-compatible server template
- Local disk writes need a volume if they must survive releases

## Common failure modes

### Health checks cannot connect
Bind the process to 0.0.0.0 and the supplied PORT.

### The process exits immediately
Read runtime logs and verify the production start file exists in the packaged release.

### A native or workspace dependency is missing
Check the lockfile, deploy root, install command, and build output in the build logs.