Railway

Railway auto-detects Node projects from package.json. A config file is optional; add one only when you want to pin the start command, set a health check, run pre-deploy migrations, or switch to a Dockerfile-based build.

When to choose Railway

You want the lowest-config push-and-it-runs experience on Node.
You want managed Postgres, Redis, or MySQL in the same project.
You like environment-per-PR with usage-based billing.

Server entrypoint

// src/server.ts
import { serve } from "@daloyjs/core/node";
import { app } from "./app.js";

serve(app, {
  port: Number(process.env.PORT ?? 3000),
  hostname: "0.0.0.0",
});

railway.json

json

{
  "$schema": "https://railway.com/railway.schema.json",
  "build": {
    "builder": "RAILPACK"
  },
  "deploy": {
    "startCommand": "node dist/server.js",
    "preDeployCommand": ["pnpm run migrate"],
    "healthcheckPath": "/healthz",
    "healthcheckTimeout": 300,
    "restartPolicyType": "ON_FAILURE"
  }
}

Or, equivalently, railway.toml:

toml

[build]
builder = "RAILPACK"

[deploy]
startCommand = "node dist/server.js"
healthcheckPath = "/healthz"
healthcheckTimeout = 300

Deploy

bash

pnpm dlx @railway/cli login
pnpm dlx @railway/cli link
pnpm dlx @railway/cli up

Gotchas

Don't override PORT. Railway injects it and the load balancer targets that port.
healthcheckTimeout is in seconds. Make it longer than your slowest legitimate startup.
Use preDeployCommand for migrations so schema changes run before traffic shifts.

Search docs