Skip to main content
Ankra billing is based on worker vCPU usage. Each plan includes a free vCPU allowance - you only pay for usage above that threshold, billed monthly. Playground environments are the exception: they are billed outside the vCPU metering. Pay-as-you-go organisations pay from credit, and saving a payment method unlocks €249 of welcome credit.

​
Plans

Ankra offers three standard plans. Organisations on a custom agreement see a dedicated plan view.
Annual commitment saves 20% on per-vCPU pricing. You can switch between monthly and annual on the Plan page.

​
Credit and payment method

Pay-as-you-go organisations pay for resources from credit. A payment method is required to use resources, but credit is always spent first, and you will never be charged without a top-up: your card is only charged when you add credit.

​
Welcome credit

Saving a payment method unlocks €249 of welcome credit, valid for 60 days. The credit is no longer granted when the organisation is created - it is added when the card is saved, from guided setup or from Billing. Each organisation receives it once, and credit you have not spent when the 60 days end expires. Until a payment method is saved, a new organisation can run its free playground, but cannot create or import other clusters. A playground on an Ankra Cloud dev cluster lasts 24 hours without a card.

​
Top-ups

Add credit from Billing → Wallet. Each top-up charges your card for the amount you choose and adds it to your balance. Automatic top-ups are off unless you turn them on. Set the balance to top up at, the amount to add, and daily and monthly caps, and Ankra adds credit when your balance drops below the threshold. We recommend turning them on for anything you need to keep running, so your services are never paused for lack of credit.

​
When credit runs out

What happens when your balance reaches zero and nothing tops it up depends on what you are running:
  • On the Ankra platform, nothing that is running is stopped. Creating new clusters and AI usage are paused until you top up. Your running clusters are unaffected.
  • On Ankra Cloud servers, including a playground that runs as an Ankra Cloud dev cluster, new resources are refused. You are notified first, and after a grace period running servers are stopped. They stay stopped until you top up.
Adding credit restores access.

​
Playground environments

A Playground is a development and evaluation Kubernetes environment on Ankra-hosted infrastructure. Playgrounds are moving to Ankra Cloud dev clusters - a single-node Kubernetes cluster managed by Ankra. This is rolling out, so your organisation may still get a playground on the shared infrastructure described below. On an Ankra Cloud dev cluster, without a payment method the free cluster lasts 24 hours. Once a payment method is saved it keeps running, paid from your credit. It is billed as Ankra Cloud usage and is not counted toward your plan’s worker-vCPU metering. The sizes below apply to playgrounds on Ankra’s shared playground infrastructure. They are sold as flat monthly list prices per size, in EUR excluding VAT, on every plan. Storage is bundled into the size. The hourly equivalent is the monthly price over a 730-hour month, for comparison only - billing is per second, not per hour. How playground charges work:
  • Pro-rata by the second. Billing runs from the moment the environment is ready (its ready_at time) until you destroy it. Nothing accrues while it is provisioning.
  • The price you order at is the price of record. Later catalogue changes apply to new orders only, never to an environment you are already running. A resize bills the old price until the resize completes and the new price after.
  • Flat price, nothing else metered. There are no per-vCPU, per-GB, egress, IP or GPU charges for playgrounds. Playground clusters are excluded from the plan’s worker-vCPU metering: they are never billed per vCPU, and they do not consume the 30 free vCPU either.
  • Own invoice line. Playground spend appears as its own line on the invoice, separate from worker-vCPU usage.
  • One free Trial per organisation. The Trial costs nothing and needs no billing plan or card. On shared playground infrastructure it expires when the organisation goes idle; as an Ankra Cloud dev cluster it lasts 24 hours unless a payment method is saved.
  • An organisation memory budget of 8 GB. Paid playgrounds share a budget of 8 GB of advertised RAM per organisation unless a limit increase has been approved. Request one from the size picker or from the Playground memory budget card on the billing page.
Ordering a paid size requires billing permissions in your organisation, a billing plan and a payment card on file. New paid orders refuse while an invoice is overdue.
Playgrounds are for development and evaluation only: no production workloads, no personal data, no backups, and they are excluded from the SLA (Terms of Service clauses 6.2 to 6.4 and 9.9).
See Sizes and Pricing on the Playground page for what each size includes and how to order, resize and destroy one.

​
Billing Overview

Navigate to Billing from the organisation sidebar to see an at-a-glance summary:
  • Current Plan - your active plan, billing cycle, commitment period, and Stripe connection status.
  • Stat cards - total vCPU, worker vCPU, billable vCPU (above free allowance), estimated monthly cost, and cost so far.
  • Cluster vCPU Summary - per-cluster breakdown of worker and total vCPU, grouped by activity. Reporting clusters show their vCPU counts with a bar indicating each cluster’s share of worker vCPU. Clusters that stopped sending node data within the last 7 days (for example because their agent is disconnected) stay listed with a Last seen status - they do not accrue usage while silent. Clusters with no node data for more than 7 days are collapsed behind a Show N inactive clusters toggle so long-gone machines don’t crowd the list.

​
Estimated Monthly Cost

The overview and analytics pages show two cost figures: The “so far” subtitle shows how much has accrued and what percentage of the month has elapsed.

​
AI Usage and Free Allowance

The overview also meters your organisation’s AI activity. Every AI request is tracked (tokens per model, priced in USD), and each plan includes a free monthly AI allowance that covers usage served on Ankra’s platform key. The overview shows month-to-date AI usage, the allowance meter, and a self-serve daily AI spend cap (USD/day) you can set as a brake on total AI activity. See Models and Cost for what the allowance covers, what pauses when it is exhausted, and how to keep unlimited AI with your own OpenRouter key.

​
Usage of every meter

Worker vCPU is one of several things Ankra measures. Usage reports every meter in one place - worker vCPU-hours, playgrounds, managed services, Ankra hosted logs and metrics, and AI - per month or day, split by cluster, application or managed service, with how completely each was measured and what it costs. Managed services and hosted logs and metrics are included during the pilot.

​
Usage Analytics

Go to Billing → Usage Analytics for detailed usage data:
  • Time range picker - view the last 24 hours, 7 days, 30 days, or a custom range.
  • Metric cards - current, peak, and average worker vCPU; active cluster count; estimated monthly cost with full-month average.
  • vCPU chart - area chart of worker vCPU over time with a dashed line for the free allowance.
  • Cost breakdown chart - hourly cost bar chart showing estimated billable cost per hour (only appears when usage exceeds the free tier).
  • Cluster breakdown table - per-cluster worker vCPU, total vCPU, and percentage share.
If all your usage is within the free vCPU allowance, the cost breakdown shows a “Within free tier” message instead of a chart.

​
Invoices

Go to Billing → Invoices to view your billing history:
  • Search and filter - find invoices by text or filter by status (all, paid, open, draft).
  • Summary cards - total invoices, combined amount and vCPU-hours for the current page, outstanding balance, and average invoice.
  • Invoice table - billing period, amount, billable vCPU-hours, status, PDF download, and a Pay Now link for open invoices.
  • Stripe portal - click “Manage in Stripe” to access Stripe’s hosted billing portal for payment methods and invoice history.

​
Changing Your Plan

Navigate to Billing → Plan & Subscription to upgrade, downgrade, or adjust your billing cycle.
1

Choose a billing cycle

Toggle between No commitment (monthly) and Annual commitment (20% discount).
2

Select a plan

Click the button on the plan you want. If the plan requires payment, you’ll be redirected to Stripe Checkout.
3

Complete payment

Enter your payment details in Stripe. After checkout you’re redirected back to Ankra and your subscription syncs automatically.
If your organisation doesn’t have Stripe connected and you select a paid plan, you’ll be prompted to set up billing first.

​
How Billing Is Calculated

Ankra tracks vCPU usage through periodic snapshots of your clusters:
  1. Snapshots are taken hourly per cluster, recording worker vCPU and total vCPU.
  2. Billable vCPU = worker vCPU minus the free allowance (per hour bucket, aggregated across clusters).
  3. Cost per hour = billable vCPU × effective rate (plan rate adjusted for any annual discount).
  4. Monthly cost = sum of hourly costs for the billing period.
Only worker node vCPU counts toward billing. Control plane nodes are not billed.

​
FAQ

Worker vCPU is the total vCPU count across all worker nodes in your clusters. Control plane / master nodes are excluded from billing.
Billing starts when your clusters report worker vCPU usage above the free allowance. If you stay within the free tier, there is no charge.
Yes. Plan changes take effect immediately. Your invoice for the month will reflect the prorated usage under each plan.
Annual billing locks in a 20% discount on the per-vCPU rate. You commit for 12 months and are billed monthly at the discounted rate. Your commitment end date is shown on the billing overview.
Usage above the free vCPU allowance is billed at your plan’s per-vCPU/hr rate. The estimated monthly cost on the overview page updates in near real-time.
The Ankra agent on that cluster has not synced node data recently, so Ankra cannot see its capacity. This usually means the agent is disconnected, the cluster was deleted outside Ankra, or node syncing is failing. A silent cluster accrues no usage. Reconnect the agent to resume metering, or delete the cluster from Ankra if it no longer exists. After 7 days without node data the cluster moves into the collapsed inactive clusters section of the summary table.

​
Next Steps

Usage Analytics

Dive deeper into vCPU trends and cost breakdowns.

Organisations

Manage teams and resources across organisations.