Skip to content

Overview: Developer Documentation

Developers can access Run:AI through various programmatic interfaces.

API Architecture

Run:AI is composed of a single, multi-tenant backend (control-plane). Each tenant can be connected to one or more GPU clusters. See Run:AI system components for detailed information.

Below is a diagram of the Run:AI API Architecture. A developer may:

  • Access the backend via the Administrator API.
  • Access any one of the GPU clusters via Researcher API.
  • Access cluster metrics via the Metrics API.

api architecture image

Administrator API

Add, delete, modify and list Run:AI meta-data objects such as Projects, Departments, Users and more.

The API is provided as REST and is accessible via the control-plan (backend) end point.

For more information see Administrator REST API.

Researcher API

Submit, delete, and list Jobs.

The API is provided as:

Researcher API is accessible via the GPU cluster itself. As such, multiple clusters may have multiple endpoints.


The same functionality is also available via the Run:AI Command-line interface. The CLI provides an alternative for automating with shell scripts.

Metrics API

Retrieve metrics from multiple GPU clusters.

See the Metrics API document.

Inference API

Deploying inference workloads is currently provided via a special inference API.

API Authentication

See REST API Authentication for information on how to authenticate REST API.

Last update: January 28, 2022