Usage & billing¶
Sistemo Cloud billing is a prepaid wallet — no subscription, no monthly fee, no surprise bills. You start with €5 free credit (no card). Adding a payment method unlocks €30 total for the trial; usage draws it down as you run; you top up when you want more. Your balance is your spend cap: at €0, new machines are blocked and running ones pause until you top up.
Cloud plan¶
| Price | Prepaid — pay only for what you run |
| Trial | €5 free on sign-up (no card); add a card (no charge) for +€25 once = €30 total trial |
| Top-up | Min €10 via Stripe (payment processor only — usage is never reported to Stripe) |
| Trial limits | 3 machines · 4 vCPU · 4 GB RAM · 30 GB storage · 2 volumes — at most 2 vCPU / 2 GB per machine |
| After your first top-up | 20 machines · 24 vCPU · 32 GB RAM · 200 GB storage · 10 volumes — at most 8 vCPU / 16 GB per machine; raised further on request |
Adding a card without topping up unlocks the €25 trial bonus but keeps the trial's limits — only a real top-up raises them.
Storage counts every disk: a machine's root disk is at least 10 GB and a volume at least 5 GB, so three trial machines use the whole 30 GB.
Limits are checked independently: filling your storage blocks both a new machine and a new volume (both need disk), while filling vCPU blocks a machine but still lets you create a volume. GET /v1/quotas shows every limit alongside what you are currently using, and each row carries a description saying what it counts.
Storage and volumes count different things
Storage is every disk on your account in any state — machine root disks and data volumes alike. Volumes counts data volumes only. So destroying a machine with ?preserve_storage=true leaves you at storage_gb 10/30 and volumes 0/2 at the same time: the kept root disk takes disk, but it does not occupy a data-volume slot. Both numbers are right.
How metering works¶
Billed per second. Quoted per hour (compute) and per GB-month (disk). Same shape as Modal and E2B: the sticker is a rate you can compare; the wallet is charged for elapsed seconds of that rate. There is no 1-hour minimum and no rounding up to the next hour.
| Metric | Clock | Sticker rate |
|---|---|---|
| vCPU | Seconds the machine is running | €0.035 / vCPU-hour |
| RAM | Seconds the machine is running | €0.012 / GiB-hour |
| Storage | Seconds the disk exists (running or stopped) | €0.050 / GB-month |
| Snapshots | Seconds the snapshot exists, on used bytes | same as storage |
Live rates: GET /v1/billing/rates and the dashboard.
A 5-second 1 vCPU / 1 GiB sandbox is 5 seconds of €0.047/hour compute, plus 5 seconds of its 10 GB disk at the GB-month rate — not a rounded hour. Idle sleep in a running machine still bills CPU + RAM; stop it for €0 compute.
Example: 1 vCPU + 1 GiB ≈ €0.047/hour while running, plus storage. €5 free ≈ 105 such machine-hours of compute (≈ 635 at the €30 trial).
Stopped is cheap
A stopped machine keeps its disk but uses no compute — you only pay storage, still prorated to the second, until you delete the volume. Stop machines you're not using instead of leaving them running.
Top-ups¶
Dashboard → Billing → top up (min €10). Card is saved for later top-ups.
What happens at €0¶
- New machines blocked and running machines pause.
- Resources kept for a grace window: 3 days on free trial, 30 days after you've topped up. Top up any time to resume.
- After grace, idle €0 account resources may be deleted (we email before this).
- Storage may nudge the balance slightly negative; cleared on next top-up.
Controlling cost¶
- Close sandboxes when done (
with Sandbox() as sb:/await sb.close()). - Stop instead of recreate if you need state between runs.
- Delete volumes and snapshots you no longer need — they bill while they exist.
- Watch Billing for live balance and usage by resource.
Viewing usage¶
Dashboard Billing: balance, top-ups, cost chart. API: GET /v1/billing/usage, GET /v1/billing/balance, GET /v1/billing/usage/by-resource (read scope).