GPU API

Submit compute jobs to our GPU cluster and monitor their status.

Current Hardware

GPU	VRAM	Count
NVIDIA GeForce RTX 5070	12 GB	1

More GPUs will be added to the cluster over time.

---

Submit a GPU Job

POST /api/gpu/jobs

Requires: Authorization: Bearer YOUR_JWT_TOKEN

Request Body

Field	Type	Required	Description
command	string	Yes	Command or script to execute
gpuType	string	No	Preferred GPU type (default: any)
dockerImage	string	No	Docker image for the job container

Example

curl -X POST https://intelligence.cognitera.ai/api/gpu/jobs \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "command": "python train.py --epochs 10 --lr 0.001",
    "dockerImage": "nvidia/cuda:12.0-base"
  }'

Response

{
  "id": "job-uuid",
  "status": "queued",
  "queuePosition": 3
}

---

Get Job Status

GET /api/gpu/jobs/:id

Requires: Authorization: Bearer YOUR_JWT_TOKEN

Response

{
  "id": "job-uuid",
  "status": "running",
  "gpuType": "NVIDIA GeForce RTX 5070",
  "gpuSecondsUsed": 145.3,
  "createdAt": "2026-03-19T12:00:00Z",
  "startedAt": "2026-03-19T12:01:00Z",
  "completedAt": null,
  "result": null,
  "errorMessage": null
}

Job Statuses

Status	Description
queued	Waiting for an available GPU
running	Currently executing on a GPU
completed	Finished successfully
failed	Execution failed (see errorMessage)
cancelled	Cancelled by user or system

---

List Jobs

GET /api/gpu/jobs

Returns the 50 most recent GPU jobs for your account.

---

Cluster Status

GET /api/gpu/status

Requires: Authorization: Bearer YOUR_JWT_TOKEN

Check real-time GPU availability.

Response

{
  "cluster": {
    "totalGpus": 1,
    "availableGpus": 1,
    "busyGpus": 0,
    "gpus": [
      {
        "name": "NVIDIA GeForce RTX 5070",
        "available": true,
        "utilization": 0,
        "memoryUsed": 512,
        "memoryTotal": 12288,
        "temperature": 42
      }
    ]
  },
  "queueDepth": 0
}