API Reference

Complete REST API reference for GPUWorker.

API Reference

All API endpoints are available at https://api.gpuworker.com. Authentication is required for all endpoints except /health, /auth/*, and /blog/*.

Authentication

POST /auth/signup

Create a new account.

{
  "name": "Ada Lovelace",
  "email": "ada@example.com",
  "password": "min8chars"
}

Response: Returns access_token and refresh_token.

POST /auth/login

Authenticate with email and password.

{
  "email": "ada@example.com",
  "password": "min8chars"
}

POST /auth/refresh

Refresh an expired access token.

{
  "refresh_token": "your-refresh-token"
}

GET /me

Returns the authenticated user’s profile. Requires Authorization: Bearer <token> header.

GPU Instances

GET /instances

List all GPU instances for the authenticated user.

POST /instances

Create a new GPU instance.

{
  "name": "my-instance",
  "gpuTypeId": "NVIDIA GeForce RTX 4090",
  "cloudType": "SECURE",
  "templateId": "pytorch-cuda12-2.1"
}

GET /instances/:id

Get details of a specific instance.

PATCH /instances/:id

Update an instance configuration.

DELETE /instances/:id

Terminate and delete an instance.

POST /instances/:id/start

Start a stopped instance.

POST /instances/:id/stop

Stop a running instance.

POST /instances/:id/restart

Restart a running instance.

Serverless Endpoints

GET /endpoints

List all serverless endpoints.

POST /endpoints

Create a new serverless endpoint.

{
  "name": "llama-inference",
  "templateId": "your-template-id",
  "gpuIds": "NVIDIA GeForce RTX 4090",
  "workersMin": 0,
  "workersMax": 3
}

GET /endpoints/:id

Get endpoint details.

PATCH /endpoints/:id

Update endpoint configuration.

DELETE /endpoints/:id

Delete a serverless endpoint.

Storage Volumes

GET /volumes

List all network storage volumes.

POST /volumes

Create a new volume.

{
  "name": "training-data",
  "size": 100,
  "dataCenterId": "US-TX-3"
}

GET /volumes/:id

Get volume details.

PATCH /volumes/:id

Update volume configuration.

DELETE /volumes/:id

Delete a volume (data is permanently lost).

Templates

GET /templates

List all container templates.

POST /templates

Create a new template.

{
  "name": "pytorch-2.1",
  "imageName": "nvcr.io/nvidia/pytorch:2.1.0-py3.10-cuda12.1",
  "dockerStartCmd": "bash -c 'jupyter lab --ip=0.0.0.0'",
  "containerDiskInGb": 20,
  "volumeInGb": 50
}

GET /templates/:id | PATCH /templates/:id | DELETE /templates/:id

Standard CRUD operations for templates.

Container Registry

GET /registry-auth

List saved registry credentials.

POST /registry-auth

Add private registry credentials.

{
  "name": "Docker Hub",
  "username": "your-username",
  "password": "your-access-token"
}

DELETE /registry-auth/:id

Remove registry credentials.

Billing

GET /billing/instances

Get GPU instance billing summary.

GET /billing/serverless

Get serverless endpoint billing summary.

GET /billing/volumes

Get storage volume billing summary.

Error Format

All errors follow a consistent format:

{
  "error": {
    "code": 400,
    "type": "bad_request",
    "message": "Description of the error"
  }
}

Common status codes: 400 Bad Request, 401 Unauthorized, 403 Forbidden, 404 Not Found, 409 Conflict, 500 Internal Error.