Provider profile
TensorDock
TensorDock is a cloud marketplace for GPU and CPU virtual machines. It suits developers who need operating-system access for model inference, training or rendering, rather than an API that only accepts prompts.
Prices and details last checked against TensorDock's own pages.
Provider details
About TensorDock
Hosting options
GPU instances use KVM virtualization and dedicated GPUs, with root access to configure the operating system and drivers. An API supports server management. Independent hosts contribute machines, so the same GPU can appear in several configurations and locations.
A virtual machine is useful when a project needs a particular driver stack, shell-based debugging or a renderer instead of a language model. It also leaves dependency installation, guest-system updates and application restarts with your team.
Prices and billing
Fund the account before deploying. Usage is deducted from the prepaid balance, and reaching zero can automatically delete servers. Individual hosts set prices; reserved arrangements are available through sales.
Where your data is hosted
The deployment dashboard is the place to choose an available host and location. A GPU model appearing in the catalog does not reserve a machine for a later deployment.
Support
TensorDock offers platform documentation and direct support. Its infrastructure service leaves guest-system configuration and the application environment with the customer.
Things to consider
Compare CPU, RAM, disk and location alongside the GPU name. Keep recoverable data outside an account balance failure, and arrange credit monitoring before leaving a job unattended.