Pricing

Per vCPU. Per GB. By the millisecond.

Top up prepaid credit and run sessions against it, on infrastructure built and run in the EU. You pay for compute only while a session is actually running — pause it and it costs storage, not compute.

€0.002 per vCPU-hour, €0.008 per GB-hour of RAM, billed per millisecond.

Tiny
€0.010/hr

1 vCPU · 1 GB RAM

Small
€0.018/hr

1 vCPU · 2 GB RAM

Medium
€0.036/hr

2 vCPU · 4 GB RAM

Large
€0.072/hr

4 vCPU · 8 GB RAM

Billed only while your code runs. Starting and stopping a session is not billed. Each execution costs at least €0.00001. Sizes are fixed tiers, priced by those two rates; custom vCPU/RAM combinations are not available.

Pause for weeks, pay cents.

A paused session keeps its disk and memory without running — that’s storage, not compute. It’s billed on the bytes it actually stores, not its disk size, at €0.03 per GiB per month: a session holding 300 MiB costs under a cent a month. Persistent volumes are billed on their provisioned size, about €0.07 per GiB per month. Captured artifacts are billed by the GB-day.

€1 free credit to start.

New accounts get €1.00 in credit, no card required. Top up from €5 whenever you need more — prepaid credit doesn’t expire.

Self-hosted

Run inis.run in your own cloud.

Some teams can’t send workloads to a shared multi-tenant fleet at all. We run inis.run inside your own cloud account, from €30,000/year.

Talk to us

Questions

What happens at zero balance?

New sessions are rejected and running work stops accruing further charges — the API returns a 402 until you top up. Nothing is deleted for running out of credit.

Do credits expire?

No. Prepaid credit doesn’t expire.

How is usage measured?

By the millisecond, only while a session is running. A paused session stops accruing compute charges and is billed for storage instead.

Can I get a custom size?

Not on the managed cloud — pick a size from the table above. Need something else, or your own cloud account entirely? See self-hosted, above.

How many sessions can I run at once?

10 concurrent live sessions per account by default, across all your API keys — 1 while your account is an unverified trial. Running against the ceiling returns a retryable 429. Need a higher limit? Ask us. Full defaults, including timeouts and disk caps, are in the session lifecycle docs.

Do you support GPUs?

No — every size is CPU-only today.

Ready to run something?

Get an API key