$0
Permanent hard-capped Compute allowance, with no usage billing or payment card. Apps pause if a limit is reached.
- 1 million requests per month
- 360 GB-hours provisioned memory per month
- 4 active vCPU-hours per month
postgres() resources. With deploy-on-push configured, branches get isolated application previews and fresh Composer-declared databases rather than production-data copies. Compute also works with external databases without Prisma ORM, but database isolation and migrations then remain the developer’s responsibility.TypeScript teams deploying APIs or full-stack apps that want application releases, Postgres provisioning, and isolated previews managed through one workflow.
Compute runs on Bun and targets HTTP services; automatic database migrations apply only to Composer postgres() resources, not external database connections.
Supported platform capabilities, execution environments, development technologies, deployment models, integrations, and triggers.
Published pricing structure, representative costs, and material billing factors.
One Prisma workspace plan includes Compute and Postgres. The base fee is paid once; these cards show Compute allowances, while Postgres operations and storage have separate meters.
$0
Permanent hard-capped Compute allowance, with no usage billing or payment card. Apps pause if a limit is reached.
$10/month
One shared workspace subscription; Compute memory, CPU and bandwidth are metered from the first unit.
$49/month
Higher request allowance with the same Compute resource rates.
$129/month
Largest published request allowance and Business workspace support; Compute resource rates remain unchanged.
Paid plans include monthly request allowances. Additional requests cost $1 per million; provisioned memory while running or kept awake, active CPU time, and outbound internet bandwidth are charged from the first unit. Idle scale-to-zero time and deployments are not billed as separate units.
| Resource | Billing Unit | Price |
|---|---|---|
| Requests beyond allowance | per million requests | $1 |
| Provisioned memory | per GB-hour while running | $0.006 |
| Active CPU | per vCPU-hour used | $0.064 |
| Outbound bandwidth | per GB sent to the internet | $0.025 |
Recurring Free workspace allowance without a credit card or usage billing. When any Free limit is reached, apps pause until the next reset or plan upgrade. Postgres has separate Free allowances in the same workspace.
Prisma offers custom pricing for higher-volume teams across its workspace products.
The monthly workspace base subscription covers Compute and Postgres once. Paid Compute usage is metered beyond requests included in the plan; memory, CPU and outbound bandwidth are billed from the first unit.
Where workloads run, how deployment is scoped, and what underlying infrastructure the platform depends on.
Prisma Compute services execute in one selected application region, not across multiple regions. Prisma Compute and Prisma Postgres share the six region choices, allowing co-location by selecting the same region for each service and database; this inventory covers Compute execution only.
Each Compute HTTP service runs in a customer-selected region. A single service cannot deploy across multiple regions.
| Location | Area | Region ID |
|---|---|---|
| San Francisco | North America | us-west-1 |
| North Virginia | North America | us-east-1 |
| Paris | Europe | eu-west-3 |
| Frankfurt | Europe | eu-central-1 |
| Tokyo | Asia Pacific | ap-northeast-1 |
| Singapore | Asia Pacific | ap-southeast-1 |
Availability reviewed · Availability source
60 seconds
Time to first response byte; streams already started are not cut off by this deadline
3 per service
Production branch only; service must have a promoted version
8 KB
Maximum per value; values are write-only and resolved at deploy time