Developer
$0 forever
For evaluating the runtime with a real endpoint.
- 2 regions
- L40S and CPU tiers
- 50 GPU-hours / month included
- Community support
- Dynamic batching and autoscaling
Get started
From zero to a live endpoint in three steps.
No cluster, no control plane, no YAML archaeology. The free tier is a real runtime with two regions and enough capacity to ship something.
- 01
Install the CLI
One binary. It authenticates with your account and talks to the control plane over HTTPS.
npm i -g @terrane/cli terrane login
- 02
Describe the runtime
A single file declares services, regions, hardware and how rollouts should behave. Commit it next to your code.
runtime: edge-inference regions: [iad, fra] autoscale: enabled batching: dynamic services: - name: embed kind: inference model: bge-m3 hardware: l40s - 03
Deploy and watch
The dry run prints the plan. The deploy streams build and rollout logs, then hands you an endpoint and the runtime view.
terrane deploy --dry-run terrane deploy terrane endpoint embed
Plans
Pay for capacity, not for seats.
Every plan runs the same runtime. The difference is regions, hardware tiers and how much of our time comes with it.
Team
$490 per month + usage
For production workloads with a small platform team.
- All 24 regions
- All hardware tiers
- Canary rollouts and warm pools
- 30-day logs, alerting
- Email support, 1 business day
Enterprise
Custom annual commitment
For regulated or high-volume environments.
- Reserved H200 capacity
- Private regions and VPC peering
- SSO, audit logs, data residency
- Named engineer, 99.99% SLA
- Migration assistance
01 Do I need Kubernetes?
No. Terrane is the runtime. You describe services in one file and the platform schedules them on managed capacity. There is no cluster to create, upgrade or right-size.
02 Can I bring my own container?
Yes. Use image: instead of model: for any service. The managed vLLM and TensorRT images are conveniences, not requirements.
03 How is usage billed?
Per GPU-second of allocated capacity plus egress, metered per region. Autoscaling to zero means idle deployments cost nothing; warm pools are billed as allocated capacity.
04 Where does my data go?
Requests are processed in the region that serves them and are not persisted beyond logs you control. Enterprise plans can pin deployments to specific jurisdictions.
05 What happens during a region outage?
The router sheds traffic from the degraded region proportionally as its health score falls, then marks it unhealthy. Neighbouring regions absorb the load. You get a notification, not a page, in most incidents.
06 Can I migrate from another inference provider?
Usually in an afternoon. terrane migrate reads common deployment formats and prints an equivalent Terrane file for review. Enterprise plans include hands-on migration help.