Preview environment compute
Charged only while an environment is building or running, for the CPU and memory allocated to your application services.
- CPU
- Memory
VeilStream system services are included and are not separately billed.
VeilStream bills preview environments for the compute they use while running, plus metered governed Postgres usage. When an environment is destroyed, compute billing stops.
First $20 of usage each month is free. No credit card required.
You pay for two categories of usage. There is no charge simply for having a project, repository, branch, or user configured.
Charged only while an environment is building or running, for the CPU and memory allocated to your application services.
VeilStream system services are included and are not separately billed.
Metered across the Postgres access you use through VeilStream. No separate proxy license.
A preview environment with 2 services, 500m CPU (0.5 core) each, 1 GiB memory each, running 4 hours uses:
…
Database usage is added separately. Image builds use the same CPU and memory rates for the minutes they take.
A concrete monthly number is easier to judge than five independent unit rates.
≈ … / month
Average per environment: …
Your usage may be higher or lower. Use the calculator below for your own workload.
These are the rates most teams pay. Query rates decrease at higher monthly volume. Higher-volume compute customers may qualify for discounted rates — talk to us.
Only pay while environments are building or running.
No separate proxy license or platform fee.
Start from a reference workload, then adjust. The estimate is runtime compute plus monthly database usage — builds add a small extra amount at the same CPU and memory rates.
Example: small application / light preview usage
Resources per service and how often you spin up preview environments.
Each deployed service normally runs as one Kubernetes pod.
Average monthly storage, traffic, and query volume across your databases.
Provisioned hosted database storage, billed as average stored gigabytes over the month.
Bytes returned to clients through the governed gateway (result traffic out).
SQL statements executed through the governed gateway.
Estimated monthly cost
…
Average per environment …
VeilStream automatically tears down preview environments when they are no longer needed. Compute billing stops with them.
That is what makes usage-based pricing fit ephemeral environments: an always-on staging namespace keeps charging; a VeilStream preview does not.
No per-seat pricing. Developers, Claude, Cursor, and automation all use the same environment and database meters — there is no separate agent or MCP fee.
Set limits on environment size and lifetime, configure spend alerts and a monthly cap, and review usage in the console before the billing period ends.
Compute billing starts when an environment begins building or running, and stops when the environment is destroyed or scaled to zero. Pull-request environments are removed when the PR closes. Branch environments follow the lifetime or idle timeout you set.
Yes. Image builds are billed at the same CPU and memory rates as running services, for the minutes the build takes. The calculator above estimates runtime; a typical build is a few extra minutes per environment.
No. There is no per-seat, per-developer, or per-agent fee. AI agents use the same pricing model as people: they consume the environment and database resources they use.
The preview environment is destroyed. Compute billing for that environment stops. Database storage continues only for databases you still have provisioned.
Each SQL statement executed through the governed gateway counts as one query.
Hosted database storage is the provisioned size of databases VeilStream holds for you, billed by average gigabytes over time. Sanitized seed snapshot files stored for restores are a separate, smaller meter.
Yes. The first $20 of usage each month is free, and you can start without a credit card. Without a payment method, $20 is also the monthly cap — deployments pause until you add a card. With a card on file, you pay only for usage above the free allowance.
Query rates decrease automatically at higher monthly volume. CPU, memory, storage, and traffic use a single published rate. For higher-volume compute, talk to us.
Yes. You can set a monthly spend cap, usage alerts, environment lifetime, idle timeout, and per-service CPU and memory limits.
Yes. The console shows current-period usage and estimated spend so you can see consumption before the month closes.
For higher-volume workloads or organizations that need additional deployment, security, invoicing, or self-hosted options, talk to us.
Talk to usStart with the free allowance and pay only for the infrastructure you actually use. First $20/month free. No credit card required.