Deploy your first server

The deploy wizard lives at Servers → Deploy. Every server boots with a public IPv4 address.

Region

Pick where the server runs. Each region lists the features it supports — block storage, floating IPs, backups. Latency to your users usually decides this.

Some regions are listed but can't be picked yet. Coming soon means we haven't opened that location; Fully booked means it's open but every machine there is taken. For either you can join the waitlist — one tap, one vote per account — and we'll know to build there next. Joining needs an account, you can leave again whenever you like, and it costs nothing. See Region waitlist.

No capacity is a third thing and it clears itself: the location is open, but no machine in it is online this second. There's nothing to put your name down for, so no waitlist appears — try again shortly.

Operating system

Choose a base image. Linux distributions log in over SSH; Windows images log in over RDP. Only images that can run in the region you picked are selectable — an image shows Coming soon where it isn't ready anywhere yet, or Not in this region when it runs elsewhere but not where you've chosen (change the region, or pick another image). The image you pick changes which login options appear in step 4.

Switch this step to One-click app to deploy with software already set up — n8n, WordPress, an AI agent, a control panel. On a plain OS you can also pick one of your saved scripts to run after boot. See One-click apps and scripts.

Size

Pick a preset plan. Each shows vCPUs, memory, SSD, and monthly transfer. Prices are monthly; the hourly rate is derived from them and capped at the monthly price. Which one you actually pay depends on your account's billing mode — monthly by default. See Billing.

Or switch to Custom and dial the hardware yourself: 1–16 vCPUs, 1–64 GB of memory in 1 GB steps, and 10–960 GB of disk in 10 GB steps, each priced per unit as you drag. The total on screen is an estimate from today's unit rates — we re-price the build when we create it and bill that figure.

A strong root password is generated the moment the page loads, so this step arrives already answered. Copy it before you move on — we don't show it again. Your options:

Root password — always works for the browser console (VNC). On Linux it also works for SSH only when the platform's guest-agent bootstrap is enabled; otherwise SSH stays key-only. On Windows it's required and named Administrator, since RDP has no key-based login.
Also add SSH keys (Linux) — layered on top rather than instead. Tick your saved keys or paste one, and save it for next time if you like. Keys you've marked auto-add arrive already ticked.
Email me the password instead — we generate one and send a secure link. It opens once and expires in 3 days. The email and the link name the account to sign in as: root on Linux, Administrator on Windows. On Linux, when the platform's guest-agent bootstrap is enabled, your first login asks you to choose a new password; on Windows nothing forces a change, so change it yourself once you're in over RDP.

Hostname

The field arrives already filled in with a readable two-word name — quiet-harbor, copper-birch, that sort of thing — so you can deploy without stopping to invent one. Type over it if you'd rather, or press the dice for a different suggestion. Letters, numbers and hyphens, up to 63 characters; the Deploy button stays disabled while the field is empty or invalid. Optionally add tags to organise and filter your fleet later.

The name is only your label, and you can change it any time. The identifier that never changes is the server's handle — see Names and handles.

You need a way in

Clear the password and select no SSH key, and Deploy switches itself off with that as the stated reason — a server nobody can log into isn't worth creating. On Windows only the password counts, because RDP has no key-based login.

Check the cost before you confirm

The wizard shows the rate for your choices and checks your wallet balance against what the deploy actually needs, which depends on your billing mode:

Monthly (the default) — the server's whole first month is charged the moment you deploy, so your balance must cover that month (after any service discount on your account).
Hourly — nothing is charged up front; your balance needs at least one hour of runway at the server's rate.

Below that, the Deploy button stays disabled and says how much you need — head to Billing and add funds, then come back.

On a monthly wallet the wizard only asks for the whole month once it has loaded your current prices and your service discount. Until then — or if either can't be loaded — it checks for an hour of runway, and the deploy itself checks the month when you confirm: if your balance can't cover it, the deploy is refused with a message saying so and nothing is charged.

If your account was set up for you by a reseller — your provider — your servers are billed to their account, not to your own wallet. The wizard says Billed to your provider's account instead of showing your balance, and your own balance doesn't hold the button back. If your provider's account can't cover the server, the deploy is refused with a message saying so; topping up your own wallet won't change that, so contact your provider.

The button also carries whichever other reason is blocking it — an image that isn't deployable yet, a custom size outside the allowed range, an invalid hostname — one at a time, in the order you'd fix them.

Deploy from code

Every deploy is a single authenticated POST /instances. The wizard shows the Equivalent API request as a ready-to-run curl snippet, so you can lift it straight into CI or a script:

curl -X POST https://console.ekire.net/api/v1/instances \
  -H "Authorization: Bearer $EKIRE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "planId": "nano",
    "regionId": "<region-id>",
    "osImageId": "<image-id>",
    "name": "quiet-harbor"
  }'

name is required. Deploying from code, you pick it yourself — the console's suggestion is a convenience of the form, not a server-side default. A custom build posts a custom object of hardware in place of planId; send exactly one of the two, and never a price — the server works that out and bills its own figure.

See the API reference for every field and the full set of endpoints.

If something goes wrong while the server is being built — for example the host can't give it the disk size you chose — the deploy fails and the server shows Error rather than starting on a smaller disk than you're paying for. A month charged at deploy is credited back; destroy the errored server and try again.

After it boots

Open the server's detail page to find its IP address, open the web console, watch live CPU/network graphs, manage snapshots and backups, resize, or reinstall the OS. If you chose the emailed-password option, open the secure link from your inbox to reveal the generated password and the account it belongs to (root on Linux, Administrator on Windows), then log in. A Linux server may ask you to choose a new password at that first login; if it doesn't — or on Windows — change it yourself.