How to Choose a LayerOne VPS Plan
Start from RAM and disk, not from a plan name. gc.pico fits the tiniest bots, gc.nano most bots, gc.micro small sites, and gc.medium is the usual production floor.
Every LayerOne plan includes KVM virtualization, Enterprise SAS SSD storage in Tampa, always-on L3/L4 DDoS mitigation, a dedicated IPv4 by default, a hypervisor firewall, and hourly billing from prepaid credit. Choose a size from the CPU, memory, and disk your workload needs. Transfer comes from your account pool, so it is not a reason to buy a larger plan.
General Compute, starting points
- gc.pico ($2.50/mo): 1 vCPU, 512 MB RAM, 20 GB SSD. Tiny bots, probes, and the lightest Linux services.
- gc.nano ($3.00/mo): 1 vCPU, 1 GB RAM, 20 GB SSD. Discord bots, WireGuard, Vaultwarden, Uptime Kuma.
- gc.micro ($5.00/mo): 1 vCPU, 2 GB RAM, 40 GB SSD. Small sites, n8n, a light panel.
- gc.small ($8.00/mo): 2 vCPU, 4 GB RAM, 60 GB SSD. Busier apps and small production.
- gc.medium ($15.00/mo): 4 vCPU, 8 GB RAM, 160 GB SSD. Coolify, Pterodactyl panel plus a node, plugin Minecraft.
- gc.large ($25.00/mo): 6 vCPU, 16 GB RAM, 320 GB SSD. Databases and multi-service hosts.
- gc.xlarge ($45.00/mo): 8 vCPU, 32 GB RAM, 500 GB SSD. Heavier production and modpacks.
Larger General Compute plans continue up the ladder. Network Optimized plans trade memory for a 10 Gbps port; no.appliance is the lightweight firewall option.
How to decide
If the bottleneck is RAM during builds or JVM heaps, step up a size. If the bottleneck is a full disk, step up or offload data. If the bottleneck is a flood of packets, DDoS is already included. Look at Network Optimized for port speed, not at a bigger General Compute plan for transfer.
Hourly credit
Plans bill hourly from prepaid account credit. Spin up gc.medium for a sprint, destroy it when the work is done, and you are not in a monthly trap. Annual credit lowers the effective rate when you already know the server will run all year. See hourly vs monthly billing.
Tampa, then test
The live region is Tampa, Florida. Use the looking glass before you migrate a latency-sensitive workload. Private-only deploy is optional when the guest should have no internet path; attach a private network for east-west traffic.
The live matrix is on the pricing page.