Skip to content
Compute

Edge Compute

Virtual machines sized to the resource, not the plan

Choose the exact vCPU, RAM and NVMe disk each server needs. It's running in under 60 seconds in any of 33 regions, billed by the hour from a prepaid balance — and bandwidth is always free.
Deploy a VMSee pricing

Free to sign up, no card needed. Top up a balance when you're ready to deploy.

Compute regions
33
From request to running
<60s
Uptime SLA
99.99%
Egress fees
$0

How it works

From blank account to root shell in a minute

No instance families to decode and no procurement. Configure a server, deploy it, and change your mind later.
  1. 01

    Configure

    Pick a region and an OS image, then set vCPU, RAM and disk independently. The console shows the exact hourly price before you commit.
  2. 02

    Deploy

    Create it from the console, the Edge CLI, the REST API or an AI agent over MCP. Your SSH keys are in place when it boots.
  3. 03

    Run, resize, stop

    Scale resources on a running VM, snapshot before risky changes, and stop it when you're done. vCPU and RAM stop billing the moment it does.

Flexible resources

Pay for resources, not instance types

Each resource has one published rate, so price scales in a straight line. Run one large server or split the same capacity across four — the bill is identical, and nothing sits idle because a plan bundled it in.
vCPU
1 – 32
RAM
1 – 64 GB
NVMe
up to 2 TB
Configuration optionsRate card

Built for production

The tools you reach for at 2am, already there

All features

  • Instant provisioning

    VMs are running in under 60 seconds. No queues, no capacity requests.

  • 33 regions

    Deploy next to your users across five continents, or in several regions for redundancy.
    See regions

  • Browser console

    A VNC console in the browser for boot problems and SSH lockouts. Rescue mode when things go really wrong.

  • Live resize

    Change vCPU, RAM and disk on a running VM. Your price changes the moment you do.

  • Private networks

    Isolated VxLAN networks between your VMs, with automatic addressing and per-customer isolation.

  • Recovery built in

    Automated backups, snapshots, OS reinstall and rescue mode — without a support ticket.

Startup scripts

Boot straight into a working stack

Attach a script at creation and it runs on first boot. Start from the Edge library or save your own, with parameters, and reuse it across every server.
  • Docker, Node.js, LAMP, LEMP, WordPress, PostgreSQL, Redis, Python and Hardened SSH
  • Custom scripts with named parameters
  • Progress and output shown on the VM page
Startup scripts docs

Live resize

Grow a server without rebuilding it

Traffic doubled? Add cores, memory or disk to the VM you already have — same IP, same data, same configuration. Billing follows the new size immediately, and the Cost Advisor tells you when a VM is oversized.
  • Resize from the console, CLI or API
  • Most changes apply without downtime
  • Scale back down just as easily
Managing VMs

Case study · Peri Happy

An AI planner backend with no ops team

Peri Happy runs its full stack on Edge. Their coding agent connects over MCP, provisions and updates the Compute instances behind the Peri Planner, and manages CDN, DNS and Shield alongside them.

  • 0

    Dedicated ops headcount

  • MCP

    Agent-managed infrastructure

Read the case studyAgent API & MCP

Operating systems

Your distro, or your own ISO

One-click images for the systems teams actually run. Need something else? Upload a custom ISO.
  • Ubuntu
  • Debian
  • AlmaLinux
  • Rocky Linux
  • CentOS Stream
  • Oracle Linux
  • Fedora
  • Alpine
  • FreeBSD
  • Custom ISO

Compare

Same VM, fewer line items

A 2 vCPU, 4 GB, 80 GB server at list price. The real gap opens up once it starts serving traffic.
All comparisons
Edge Compute compared with Vultr, DigitalOcean and AWS EC2
VultrDigitalOceanAWS EC2
2 vCPU · 4 GB · 80 GB / month$19.24$20$24~$36
EgressFreeAllowance, then per GBAllowance, then per GB$0.09/GB
Size each resource independently
Sign up without a card
Prepaid — no surprise bills
DDoS protectionIncludedExtra costBasicExtra cost

Competitor figures are published list prices at the time of writing; AWS is an estimate for t3-family on-demand in us-east-1 plus gp3 storage. See each comparison for detail.

FAQ

Before you deploy

Something else? Ask the compute team.
How is Edge Compute billed?
Per resource, by the hour, from a prepaid balance: $0.004 per vCPU-hour, $0.00256 per GiB of RAM per hour, and $0.074 per GiB of NVMe per month. Bandwidth is free. There's no subscription to cancel.
Do I need a card to sign up?
No. Creating an account is free and doesn't ask for a card. To run VMs you load a prepaid balance — and a one-time $5 starter credit is available once a card is on file.
What happens if my balance runs out?
Running VMs stop; nothing is deleted. Disks, configuration and IPs stay in place, and the VMs we stopped start again automatically when you top up. Turn on auto top-up to avoid interruptions.
Are the vCPUs shared?
Each vCPU is a dedicated virtual core — no noisy neighbours competing for your CPU time or memory. If you need whole machines to yourself, look at private clusters.
Can I run GPU workloads?
Not yet. GPU instances are coming soon — join the waitlist on the GPUs page and we'll let you know, or talk to us about early access.

Guides

Plan, deploy and run it well

Your next server, running in a minute

Configure exactly what you need, see the price before you deploy, and pay only while it runs. Bandwidth is on us.

Deploy a VMTalk to the compute team

Free signup, no card. Prepaid billing — a runaway bill is impossible by design.