TIR

TIR is E2E Cloud’s AI-native platform: GPU and CPU containers and VMs, Kubernetes and Slurm clusters, inference endpoints, and fast shared storage for training and serving AI.

Open TIR

Sign up

Getting Started

Set Up a ProjectA project with a teammate in it, ready for your first resources.Launch a GPU InstanceA GPU container with JupyterLab and SSH that deploys in seconds.Launch a VMA Linux VM with root access over SSH, billed by the minute only when you use it.Deploy a ModelA managed, dedicated endpoint for Hugging Face models or your own custom models.

Compute

InstancesGPU or CPU containers with JupyterLab, SSH, and ready-made images.Virtual MachinesLinux VMs on GPU or CPU plans, with their own kernel and root access over SSH.Load BalancerOne address that spreads web traffic across your VMs and Auto Scaling Groups.Auto Scaling GroupsGroups of identical VMs that grow and shrink with demand or on a schedule.KubernetesHighly available clusters with a three-node control plane and worker pools that autoscale.

Training and Inference

Slurm ClusterGPU nodes with a managed Slurm scheduler for multi-node training.Private ClusterDedicated GPU nodes your account owns, shared across its projects.Model EndpointsModels served behind one HTTPS address, with autoscaling, logs, and monitoring.

Storage and Databases

Block StorageDisks of the size you choose, mounted on an instance or VM like a local disk.Object StorageS3-compatible buckets that instances mount and S3 tools read and write.Shared File SystemShared read-write storage that many instances and jobs mount at once.Parallel File SystemFast parallel storage shared by instances and training clusters.DBaaSManaged PostgreSQL with snapshots, an IP allowlist, and settings you can tune.

Networking

VPCA private network in your project, with an IPv4 range you choose, for CPU VMs.Reserve IPA public or private IP that stays yours until you delete it.Security GroupsInbound and outbound traffic rules for instances, endpoints, and Slurm clusters.

Platform Services

PipelinesRun multi-step Argo Workflows or Kubeflow Pipelines on a CPU or GPU plan, once or on a schedule.Container RegistryA private Docker image registry in each project, used by instances and endpoints.Alert ManagementAn email, Slack message, or webhook when a metric on a resource crosses a threshold.

Account and Billing

ProjectsSeparate workspaces, each with its own members, policies, usage, and activity.IAMUsers, roles, and policies that decide who can do what in your account.Billing and UsageCredits, coupons, invoices, transactions, and usage for your account.SupportSupport tickets for cloud, billing, and sales questions, from raising to closing.

Developer Tools

TIR API ReferenceEvery TIR API operation, with sample requests in cURL, Python, and JavaScript.Personal Access TokensTokens your scripts and API calls use to sign in, in place of your password.SSH KeysKeys that replace passwords when you sign in to instances and training clusters.