Pricing

Predictable cloud pricing.
Review your price before you deploy.

One price for the resources you reserve. For an unchanged profile, the price doesn't move with traffic. Start with the free $5 starter credit.

No ingress or egress fees No per-invocation charges Price shown before you commit

Static sites

Free

Landing pages, docs sites, marketing sites, and SPAs are free for verified accounts. No payment card required.

  • Up to 3 active static sites per account
  • 1 GB total pinned assets included
  • Verified account required. No card required.
Managed Cloud

Four presets. One reserved deployment each.

Web apps, APIs, and agents run on fixed resource presets. Each is priced up to a monthly maximum for that reserved deployment, prorated by running time.

Starter

Smallest
$7/ month max

Up to $7 per month for this reserved deployment, prorated by running time.

  • 0.5 vCPU
  • 1 GB RAM
  • 2 GB ephemeral storage
  • Shared endpoint included

Growth

Most common
$14/ month max

Up to $14 per month for this reserved deployment, prorated by running time.

  • 1 vCPU
  • 2 GB RAM
  • 5 GB ephemeral storage
  • Shared endpoint included

Scale

Production
$28/ month max

Up to $28 per month for this reserved deployment, prorated by running time.

  • 2 vCPU
  • 4 GB RAM
  • 25 GB ephemeral storage
  • Shared endpoint included

Pro

Heavy
$56/ month max

Up to $56 per month for this reserved deployment, prorated by running time.

  • 4 vCPU
  • 8 GB RAM
  • 50 GB ephemeral storage
  • Shared endpoint included

Attached services (Postgres, Redis, MongoDB, MySQL, MinIO) are separate Managed Cloud resources: each selects its own preset and bills as another reserved deployment. Persistent NVMe storage is an add-on at $0.16 per GB-month. Replicas reserve separate resources: 2 replicas cost 2x the preset.

Start with the free $5 starter credit, valid 14 days on Managed Cloud deployments with a valid payment method.

CPU VMs

Hourly billing with a hard monthly cap.

Every VM includes a 20 GB NVMe boot disk and one public IPv4 address. No separate ingress or egress charge. Available as regions pass verification.

Shared CPU

Shared vCPU cores
SizevCPURAMHourlyMonthly cap
Shared S22 GB$0.0192/hr$14/mo
Shared M34 GB$0.0260/hr$19/mo
Shared L48 GB$0.0384/hr$28/mo
Shared XL816 GB$0.0671/hr$49/mo
Shared 2XL1632 GB$0.1219/hr$89/mo

Dedicated CPU

Dedicated vCPU cores
SizevCPURAMHourlyMonthly cap
Performance S24 GB$0.0384/hr$28/mo
Performance M48 GB$0.0671/hr$49/mo
Performance L816 GB$0.1219/hr$89/mo
Performance L-Memory832 GB$0.2055/hr$150/mo
Performance XL1632 GB$0.2329/hr$170/mo
Performance 242448 GB$0.3438/hr$251/mo
Performance 323264 GB$0.4534/hr$331/mo
Performance 6464128 GB$0.8973/hr$655/mo

Additional NVMe storage: $0.09 per GB-month, prorated hourly. Billing continues while the VM's resources remain reserved, until the VM is deleted. Prices are flat across admitted regions.

GPU

LLM inference, generative models, and ML workloads on RTX 3090, 4090, 5090, L40s, A100, H100, and H200 class hardware.

GPU pricing is quoted live per hour from currently available inventory. The exact quote is shown at selection, before you commit. GPU workloads are paid-only and do not consume the starter credit.

AI Gateway

One endpoint for hosted AI models, billed from a prepaid balance.

  • Provider cost + 5% Varity fee
  • $20 minimum refill, prepaid balance
  • Valid payment method required

AI Gateway balances are separate from Cloud, VM, GPU, and static products.

Common questions

Everything builders ask about predictable pricing.

How does Varity pricing work?+

Varity prices the resource profile you reserve, not your traffic. Static sites are free within the free-tier limits. Managed Cloud apps use fixed presets billed up to a monthly maximum, prorated by running time. CPU VMs bill hourly with a hard monthly cap. GPUs are quoted live at selection. The exact price is shown before you deploy.

Will my bill go up if my app gets popular?+

For an unchanged resource profile, traffic alone never creates a usage charge. There are no bandwidth overages, no per-invocation charges, and no build-minute caps. Changing your resources, adding services or replicas, adding accelerators, or adding apps can change the price, and any change is shown before you confirm it.

Is there a free tier?+

Yes. Static sites are free for verified accounts: up to 3 active static sites and 1 GB of total pinned assets per account. No payment card is required for the free static tier. For dynamic apps, you can start with the free $5 starter credit on Managed Cloud.

How does Managed Cloud billing work?+

Each preset is priced as up to a fixed amount per month for that reserved deployment, prorated by running time down to the minute. If your deployment runs the whole month, you pay the monthly maximum; if it runs half the month, you pay about half. Attached services like Postgres or Redis are separate Managed Cloud resources with their own presets.

How are CPU VMs billed?+

Hourly, with a hard monthly cap per VM. Every VM includes a 20 GB NVMe boot disk and one public IPv4 address, with no separate ingress or egress charge. Additional NVMe storage is $0.09 per GB-month, prorated hourly. Billing continues while the VM's resources remain reserved, until you delete the VM.

What happens if I outgrow my resource profile?+

Move to a larger preset or VM size. You pick the new profile, review the new price before confirming, and that is what you pay from then on. No overage charges and no traffic-spike billing events between profile changes.

See your exact price.

Pick a profile at the developer portal and the exact price is shown before you deploy. That is the number on your bill.