We use cookies to improve your experience and analyze site traffic.

Product · Compute

Run your code. We do the rest.

Push to GitHub or run a CLI command. We detect your framework, install the deps, build the image, run health checks, place it behind a load balancer, and give you a URL with automatic SSL. Sixty seconds, end to end.

Auto-detected runtimes
Node 20Python 3.12Go 1.22Ruby 3.3BunDenoJava 21+ Dockerfile
starship · api · deploy
deploying
Deploying v1242 / 6 stages
  1. Detected Node 20
    package.json · express · pg
  2. Pulled cached deps
    143 packages · 2.1s
  3. Built image
    147 MB · pushed to registry
  4. Booted instance
    us-east-1 · 0.5 vCPU · 1 GB
  5. Health check ok
    GET /healthz · 200 · 12 ms
  6. Routed traffic
    api.starship.com · auto SSL
building…
package.json
{
··"name": "starship-api",
··"engines": { "node": "20.x" },
··"scripts": {
····"build": "tsc", ← detected
····"start": "node dist/index.js", ← detected
··},
··"dependencies": {
····"express": "^4.19.0" web
··}
}
runtimeNode 20
frameworkExpress
installnpm ci
buildnpm run build
startnpm start
portauto-detected
Push to deploy

We read your repo.

Drop us at any Node, Python, Go, Ruby, or Bun project and we figure out how to run it. We pick the runtime version, install the right package manager, run your build, and pick a healthy starting size. No Dockerfile required. If you prefer one, we'll respect it.

starship · api · resize
live · zero-downtime
Pick a sizecurrent · Medium
metered every minute · pay only for what's allocatedapply →
Resize anytime

Pick a size. Change your mind.

Right-size from 0.25 vCPU all the way up. Bump it later and the change rolls out without dropping a request. Pay only for what's actually allocated, metered every minute.

starship · api · copies
load-balanced · 3 of 3 healthy
copies
load balancer
copy 1
copy 2
copy 3
3 copies · traffic split round-robinhealthy · serving
Need more capacity

Add a copy. Get a load balancer.

Run more than one copy of your service and traffic spreads across all of them automatically. If one stops responding, the next request goes to a healthy one. Add or remove copies in one click. The load balancer is on the house.

the bill

Pay for what you allocate. Nothing more.

Compute is metered every minute against the CPU and memory allocated to your service. No per-seat fees, no minimums. Every account starts with $5 in credits.

CPU
$0.0596per vCPU-hour
Graviton (ARM64)
Memory
$0.0065per GB-hour
allocated to your service

Deploy your first service. Sixty seconds.

No DevOps required. No Dockerfile required. No credit card required.