Private Cluster
TIR Private Cluster gives you a dedicated pool of GPU-backed compute nodes shared across multiple teams and projects. Fixed billing, guaranteed capacity, and role-based access control — designed for enterprise AI workloads at scale.
Quick Start
Create Your First Cluster
Reserve a dedicated GPU pool with fixed pricing and guaranteed capacity.
Private Cluster Features
Node states, monitoring, RBAC, and multi-project GPU sharing.
Manage Allocations
Allocate, deallocate, and monitor GPU nodes across projects.
Troubleshooting & FAQs
Resolve common issues and get answers to frequently asked questions.
Explore Private Cluster
Node Allocation
Distribute GPUs across projects
Access Control
IAM roles & permissions
Monitoring
Node health & resource metrics
API Reference
Private Cluster API Reference
Programmatically create, configure, and manage TIR Private Clusters. Automate node allocation, retrieve cluster status, and control GPU distribution via REST.
/teams/{id}/private-clustersList all private clusters/teams/{id}/private-clustersCreate a private cluster/teams/{id}/private-clusters/{cluster_id}Get cluster details/teams/{id}/private-clusters/{cluster_id}Update cluster configuration/teams/{id}/private-clusters/{cluster_id}/allocateAllocate nodes to a project/teams/{id}/private-clusters/{cluster_id}Delete a private clusterBilling & Plans
Billing & Credits
Private Clusters are billed at a fixed rate per node, independent of GPU utilization. No additional charges for workloads deployed inside the cluster.
Fixed-rate billing
Billed per node per hour regardless of utilization — predictable costs at scale.
No per-workload charges
Deploy Nodes, Inference Endpoints, Training Clusters, and Vector DBs at no extra cost.
Committed plans
Reserve nodes for 30, 90, or 365 days and get a lower rate compared to hourly pricing.