Linux servers, ready in minutes.
Self-managed Linux servers on AWS. You get root and run what you like; Rig runs the machine, the network, the backups and the panel. This page covers everything a server comes with, and the rules behind it.
Four sizes, three shapes.
Every size has the same features. Billed by the hour, never more than the monthly price; or yearly with two months free.
| Size | CPU and memory | Disk | Traffic a month | Backups | Price | |
|---|---|---|---|---|---|---|
| Basic | 1 vCPU · 2 GB | 40 GB SSD | 2 TB | Weekly, 1 save point | … | Choose Basic |
| Good | 2 vCPU · 4 GB | 80 GB SSD | 3 TB | Daily, 3 save points | … | Choose Good |
| Epic | 4 vCPU · 8 GB | 160 GB SSD | 5 TB | Daily, 5 save points | … | Choose Epic |
| Epic Plus | 8 vCPU · 16 GB | 320 GB SSD | 6 TB | Daily, 5 save points | … | Join the waitlist |
- GeneralBalanced CPU and memory.
- CPU-optimisedTwice the vCPUs, for builds, workers and compute.
- Memory-optimisedTwice the memory, for databases and caches.
What you choose, and what happens next
Four choices, then a few minutes
Pick the size and shape, the region, the operating system (an OS image, an app stack, or Rig's pick) and a name. Everything else has a sensible default, under Advanced:
- Hostname, and how you connect over SSH
- A private network and a firewall, new or one you already have
- A shared or dedicated IPv4, and how long backups are kept
- A startup script of up to 64 KB that runs once on first boot
coffee-club · Good · General · London
Payment receivedDoneServer createdDoneFirewall and networkDoneOperating systemDoneSSHDoneBackupsDoneReady at 203.0.113.24coffee-club.rig.app
Open the panel, copy the SSH command, or ask The Duke to deploy.
Your email and card are checked before anything is created, and pressing the button twice never makes two servers. If setup fails after you pay, Rig retries twice, then refunds you automatically.
Connect the way you like
One choice, change it any time
Password login is always off. Whatever you pick, the web console works, so nobody is ever locked out.
- Keys pasted in, imported from a GitHub account, or saved on your account for every server
- ed25519, ECDSA, or RSA of 3072 bits or more
- Key changes reach the server in seconds; removing a key removes it everywhere
- The root password is never emailed: a reset shows it once
| Option | What it means |
|---|---|
| Through Rig | Connect from Rig on your Mac. No keys to manage, and port 22 stays closed to the internet. |
| With my SSH key | Normal SSH with your own key, from anywhere. |
| Both | The two together. |
Addresses, domains and the firewall
Online with HTTPS from the first minute
Each server has a site address such as coffee-club.rig.app with HTTPS, a shared IPv4 and its own IPv6.
- Custom domains, with HTTPS issued automatically and help with DNS
- Firewall rules in and out, reusable across servers; removing the last SSH rule warns you first
- Private networks between your servers
- A load balancer on Epic, included in Epic Plus
| Address | Reverse DNS |
|---|---|
| Shared IPv4 | Rig's, fixed |
| Your IPv6 | You set it |
| Dedicated IPv4 (add-on) | You set it, once the name points back at the IP |
Port 25 is closed on new servers to stop spam. Mail over port 587 or an API works straight away, or ask in Networking to open 25.
Undo, restore, grow
Backups and save points
Backups run at a time you set and are kept 7, 14 or 30 days. Save points are snapshots you take before a risky change.
- Back up now, any time
- Restore in place: a save point is taken first, so a restore can be undone
- Start a new server from any backup, or download it
- Delete and rebuild protection, per server
Resizing
Basic to Good to Epic to Epic Plus, or a custom size. The server restarts, which takes about a minute, and a save point is taken first.
- Change CPU and memory only, so you can size back down later
- Or grow the disk too; a disk can grow but never shrink
- Now or at a time you schedule
- The price change is shown before you confirm
Know before your users do
Alerts with sensible defaults
Alerts go by email, in the app and to Slack. Change any threshold per server.
- Uptime checks on the site address every minute from 3 locations: an alert after 2 failures in a row, and another when it recovers
- Spend alerts: an email when this month's bill passes an amount you set
- Maintenance notices at least 7 days ahead, except for security fixes
- Rig Launch and each region on the public status page
| Alert | Default |
|---|---|
| CPU | Over 90% for 10 minutes |
| Memory | Over 90% |
| Disk | Over 90% |
| Down | For 2 minutes |
| Uptime guarantee | 99.9% a month, credits applied automatically |
Staff stay out. Unless you let them in.
Grant support access for the servers, level and time you choose: 1 hour, 24 hours or 7 days. A banner shows while it is on, revoking is one click, and you are emailed when it starts, when staff first use it and when it ends.
| Level | What support can do |
|---|---|
| View | See status, metrics, logs and config. No shell. |
| Operate | Also reboot, power on and off, restore backups and change the firewall. |
| Shell | Also use the console and SSH as root, with short-lived credentials made for that session. |
Without a grant, staff see only billing, the server's size, region, status and IPs, and host-level usage. For abuse or a legal order they can power a server off but never open it, and you are emailed the reason at once.
Who can do what
| Action | Owner | Operator | Member | Viewer |
|---|---|---|---|---|
| See servers and their status | Yes | Yes | Yes | Yes |
| Power, reboot, console, SSH keys, firewall, backups, save points | Yes | Yes | No | No |
| Order, resize, add-ons, rebuild, delete, transfer | Yes | No | No | No |
| Grant support access | Yes | No | No | No |
Team servers belong to the team and the owner pays. Each member adds their own SSH keys, and removing someone from the team removes their keys from its servers. Group servers into projects and tags.
What happens when
| When | What happens |
|---|---|
| The disk fills up | The server stays up and writes fail. You're warned at 80% and 95% and offered more disk or a bigger size. |
| You reach your traffic | The site is slowed, not stopped, and never charged extra unless you turned on auto top-up. |
| A payment fails | Servers keep running for 7 days, then are powered off with everything kept. At 30 days a final backup is kept for 14 days. |
| You delete a server | Undo it from the panel until the deletion runs. Its final backup is kept for 14 days. |
| You move region | It takes about 10 minutes and the shared IP changes; you're warned to update DNS first. |
| The provider has an outage | The server is marked degraded on the status page and staff are alerted. |
Script it, or let your agent do it
Everything the panel does, from code
A REST API with scoped tokens, rig launch commands in the Rig CLI, a Terraform provider, and Rig's MCP tools for AI agents.
- CLI: create, list, ssh, power, resize and backup
- Terraform: servers, firewalls, volumes and more
- MCP writes wait for your approval in Rig, like every Rig MCP write
rig launch ssh coffee-clubKeep going
- Overview Plans, prices and how a deploy goes.
- The Duke Rig's AI agent: deploys, code review, failures and setup, each step after you approve it.
- Apps Ship from Rig or Git without a server, or run managed Kubernetes.
- Databases Managed Postgres and Redis, started from the database already on your Mac.
- Storage Extra disks, S3-compatible buckets with a CDN, your own images, and free DNS.
- GPU NVIDIA L4 GPU servers by the hour, for models, media and batch jobs.