How Pricing Works
The four ways a Cosmoner resource can be priced, and the billing rules that apply to all of them.
This section lists the rates for each product and explains how each one is measured and charged.
Where a product has published rates, the table on its page is generated from the same price catalogue the control panel charges against, so it cannot fall behind what you are actually billed. Two products have no table, because they have no single rate to publish: servers are priced per size and domains per extension, and both show you the price before you commit to one.
For invoices, payment methods, credits, and what happens when a payment fails, see Billing & Payments.
The four pricing models
Every resource uses one of four models. Knowing which one a product uses tells you what changes your bill.
| Model | What you pay for | Products |
|---|---|---|
| Flat | A fixed monthly price for the resource existing | Static sites, container registry base fee, secrets packs |
| Per size | A monthly price attached to the size you picked | Servers, web service instances, managed Redis plans |
| Tiered | A prepaid allowance bundle; the tier price is the whole bill until you exceed it | Object storage |
| Metered | A rate multiplied by measured usage | Registry storage and egress, CDN delivery, free-plan email |
A single product can combine them. A container registry, for example, has a flat monthly fee and metered storage and egress.
Rules that apply to everything
You are charged from the day you create it
Every resource is prorated from the moment it is provisioned to the next billing date. Deploying mid-month costs the fraction of the month you actually get, and the full monthly price applies from the next cycle onwards.
The prorated amount is shown before you confirm — in the create flow for servers, apps, buckets, databases, and registries — so the charge is never a surprise.
Removing a resource credits the unused days
Deleting a resource stops its billing immediately and adds the unused portion of the current period to your project's credit balance, which is applied automatically to upcoming invoices. This happens whether or not you still have other resources running.
Credits and refunds are covered in full under Removing Resources and Credits.
Stopping is not the same as deleting
A stopped server still reserves its disk and its IP address, so it is still billed. Only deleting the resource stops the charge. The same applies to any resource that can be paused rather than removed.
Allowances reset with the billing period
Where a product includes an allowance — object storage tiers, registry egress, email quotas — the allowance is per billing period and does not roll over. An unused allowance is not credited.
Usage is measured per organization, not per resource
Allowances that span resources — registry egress is the clearest case — are totalled across your whole organization for the period, not counted separately for each registry.
Per-product pages
- Apps — instance sizes, static sites, and the egress IP add-on
- Servers — per-size pricing and what stopping costs
- Object Storage — how the three allowances interact, and what happens over the limit
- Container Registry — how stored bytes and pulls are measured
- Redis — plan families, and what high availability changes
- Email — quota, daily and hourly ceilings, and pay-as-you-go
- Secrets — the free allowance and how packs are added
- Domains — registration, renewal, and what is included