High Frequency cores, fast network storage,
intrusion protection and off-site backups. From €2 a month.
€25 free credit to try anything. No credit card required.
Connect to a live server. No account needed.
Production hardware, enforced limits. 15-minute sessions.
Like what you see?
Get your own instanceHere they come standard, on every account.
A fire, a hack or a blocked account cannot reach your backups: they live at a different provider, in a different country and jurisdiction, encrypted with your key before they leave the hypervisor.
Warden inserts a traffic inspection point in front of a resource: reputation filtering against curated threat lists, intrusion detection with automatic bans, inline exploit dropping at the top tier. It watches both directions, including traffic leaving a compromised workload.
Every disk is a replicated network disk that behaves like a local RAID: a dead host loses nothing, and the speed is still there. Around 160 µs latency, up to 200k IOPS and 4 GB/s per disk.
High Frequency VMs run AMD EPYC 4565P at up to 5.7 GHz: Geekbench 6 single-core 3336, about 3x a typical cloud vCPU. From €5 a month.
Pick what you need. See what you pay. No surprises.
Every VM, disk and port carries written limits, and the platform delivers exactly those numbers. Here is how each resource behaves, including at the edge.
Full speed by default. When a host runs hot, Burst VMs yield first, down to a written floor each class carries in the catalog. You can read the floor before you buy.
Top scheduling priority: full core speed under any host load, around the clock. Switching a VM between profiles applies live.
A Burst VM pegging its CPU for minutes is stepped down gently, ten percent at a time, and restored to full speed the moment it goes idle or the host cools off.
A 5,000 IOPS tier delivered 5,016 in testing, around the clock. Limits are enforced by the hypervisor per disk, so a neighbor physically cannot spend your budget.
Speed tiers from 5,000 to 200,000 IOPS, plus power-loss-protected classes for databases. A tier change applies in place, with the data staying put.
IO queues: latency rises while requests keep completing. The per-disk metrics chart your usage against the limit line, so you see it before your users do.
Every VM gets the same unthrottled multi-gigabit port on a flat high-speed fabric. Between your VMs the fabric tops out around 200 Gbit/s, hardware permitting.
5 TB of external and 50 TB of internal traffic per VM each month, with inbound always free. The allowance covers most workloads entirely.
Traffic keeps flowing past the allowance and overage bills at a flat published rate per terabyte. The bill grows linearly and you can see it coming.
For readers who want to know how it actually works: the cluster design, the failure domains, the hardware underneath.
Every physical node runs compute, storage, and network, wired together like an HPC machine: storage moves over a direct memory path between nodes, and the fabric runs at interconnect speeds up to 200 Gbit/s. Data is replicated across nodes with configurable redundancy. Lose a node, lose two nodes, the cluster keeps running. Your VMs migrate automatically, storage stays consistent, network keeps routing.
Storage isn't a SAN in the basement. Network isn't a switch with VLANs. Both are software layers that follow the workload. Need to move a VM? Storage follows. Need to resize a disk? Done live. Need maintenance? VMs evacuate, we swap hardware, VMs return. You notice nothing.
We don't run one giant cluster. Infrastructure is split into isolated cells, each a self-contained failure domain. A bug in one cell, a bad firmware update, a network misconfiguration stays contained. Your workload in cell A doesn't care what happens in cell B.
Control panel, node agents, orchestration layer, billing engine, monitoring, API: the entire stack is written in-house, top to bottom. When something needs fixing, we fix it in hours, because every line of it is ours.
Failure is a design input here. What each scenario actually does to your workload:
We showed you the specs. Now see if they hold up.
The smallest VM is €2.50 a month all-in, and the first 25 GB of disk are free on every VM.
