> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.ibee.co.in/docs/api-reference/gpu-vms/create-gpu-vm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ibee.co.in/_mcp/server.

# Create a GPU VM

POST https://api.ibee.co.in/v1/compute/gpu-vms
Content-Type: application/json

Creates a GPU VM. Returns an operation you can poll for status. Requires scope: vm.write.

Reference: https://docs.ibee.co.in/docs/api-reference/gpu-vms/create-gpu-vm

## Authentication

- `Authorization` header (bearer token, required) — IBEE platform API token. Production tokens start with `ibee_prod_key_`; development tokens start with `ibee_dev_key_`. Use a token only with its matching gateway environment.

## Servers

- `https://api.ibee.co.in/v1` (Production, default)
- `https://api.ibee.co.in/v1` (Development)

## Request

### Query parameters

- `workspace_id` (string, required) — The positive numeric workspace ID to scope this request to. Zero, negative, and non-numeric values are rejected.

### Headers

- `X-Idempotency-Key` (string, required) — Unique key used to safely retry write operations.

### Body (application/json)

This endpoint expects a CreateGpuVmRequest.

- `name` (string, required) — Display name for the virtual machine.
- `os_distro` (string, required) — Operating system distribution (e.g. ubuntu, centos, debian, rocky).
- `os_type` (enum, required) — Operating system family.
  - Allowed values: `linux`
- `template_id` (string, required) — GPU-compatible template ID returned by the compute catalog.
- `cpu` (integer, required) — Number of vCPUs.
- `ram_mb` (integer, required) — RAM in megabytes.
- `gpu_count` (integer, required) — Number of GPUs to attach.
- `gpu_model` (string, required) — GPU model (e.g. A100, H100, L40S, RTX4090).
- `plan_id` (string, required) — Billable GPU plan ID returned by the compute catalog.
- `billing_catalog` (map from string to any, required) — The selected plan's `billing_catalog` object from `GET /compute/plans?vm_type=gpu`, passed through unmodified — the same flow the portal uses. Treat it as opaque. Billing charges are always computed server-side from `plan_id`, not from this echoed copy.
- `site_id` (string, optional, nullable) — Optional placement site ID. Omit for automatic placement. To pin the VM, copy `site_id` from `GET /compute/sites` and use the same value when filtering plans and images.
- `disk_gb` (integer, optional, nullable) — Root disk size in gigabytes.
- `ssh_key_ids` (list of string, optional) — SSH key IDs to inject into the VM.
- `tags` (list of string, optional) — Arbitrary tags for filtering and organization.

## Response

### 202

VM creation accepted.

- `operation_id` (string, required)
- `vm_id` (string, required)
- `status` (string, required)
- `submitted_at` (datetime, required)

## Errors

### 400 Bad Request Error

The request is invalid.

- `detail` (ErrorDetail, optional)

### 401 Unauthorized Error

The API token is missing or invalid.

- `detail` (ErrorDetail, optional)

### 402 Payment Required Error

Billing denied the resource creation request. No resource is provisioned; inspect the error reason, resolve billing, and retry with the same idempotency key where supported.

- `detail` (ErrorDetail, optional)

### 403 Forbidden Error

The token is not allowed to access the workspace or does not have the required scope.

- `detail` (ErrorDetail, optional)

### 409 Conflict Error

The request conflicts with the current resource state.

- `detail` (ErrorDetail, optional)

## Types

### ErrorDetail

## Examples

**Request**

```json
{
  "name": "ml-training-01",
  "os_distro": "ubuntu",
  "os_type": "linux",
  "template_id": "tmpl_ubuntu_2204_cuda",
  "cpu": 8,
  "ram_mb": 32768,
  "gpu_count": 1,
  "gpu_model": "A100",
  "plan_id": "string"
}
```

**Response**

```json
{
  "operation_id": "op_123",
  "vm_id": "vm_abc123",
  "status": "accepted",
  "submitted_at": "2024-01-15T09:30:00Z"
}
```

**SDK Code**

```python
import requests

url = "https://api.ibee.co.in/v1/compute/gpu-vms"

querystring = {"workspace_id":"710995"}

payload = {
    "name": "ml-training-01",
    "os_distro": "ubuntu",
    "os_type": "linux",
    "template_id": "tmpl_ubuntu_2204_cuda",
    "cpu": 8,
    "ram_mb": 32768,
    "gpu_count": 1,
    "gpu_model": "A100",
    "plan_id": "string"
}
headers = {
    "X-Idempotency-Key": "X-Idempotency-Key",
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers, params=querystring)

print(response.json())
```

```javascript
const url = 'https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995';
const options = {
  method: 'POST',
  headers: {
    'X-Idempotency-Key': 'X-Idempotency-Key',
    Authorization: 'Bearer <token>',
    'Content-Type': 'application/json'
  },
  body: '{"name":"ml-training-01","os_distro":"ubuntu","os_type":"linux","template_id":"tmpl_ubuntu_2204_cuda","cpu":8,"ram_mb":32768,"gpu_count":1,"gpu_model":"A100","plan_id":"string"}'
};

try {
  const response = await fetch(url, options);
  const data = await response.json();
  console.log(data);
} catch (error) {
  console.error(error);
}
```

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995"

	payload := strings.NewReader("{\n  \"name\": \"ml-training-01\",\n  \"os_distro\": \"ubuntu\",\n  \"os_type\": \"linux\",\n  \"template_id\": \"tmpl_ubuntu_2204_cuda\",\n  \"cpu\": 8,\n  \"ram_mb\": 32768,\n  \"gpu_count\": 1,\n  \"gpu_model\": \"A100\",\n  \"plan_id\": \"string\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("X-Idempotency-Key", "X-Idempotency-Key")
	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(res)
	fmt.Println(string(body))

}
```

```ruby
require 'uri'
require 'net/http'

url = URI("https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["X-Idempotency-Key"] = 'X-Idempotency-Key'
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"name\": \"ml-training-01\",\n  \"os_distro\": \"ubuntu\",\n  \"os_type\": \"linux\",\n  \"template_id\": \"tmpl_ubuntu_2204_cuda\",\n  \"cpu\": 8,\n  \"ram_mb\": 32768,\n  \"gpu_count\": 1,\n  \"gpu_model\": \"A100\",\n  \"plan_id\": \"string\"\n}"

response = http.request(request)
puts response.read_body
```

```java
import com.mashape.unirest.http.HttpResponse;
import com.mashape.unirest.http.Unirest;

HttpResponse<String> response = Unirest.post("https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995")
  .header("X-Idempotency-Key", "X-Idempotency-Key")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"name\": \"ml-training-01\",\n  \"os_distro\": \"ubuntu\",\n  \"os_type\": \"linux\",\n  \"template_id\": \"tmpl_ubuntu_2204_cuda\",\n  \"cpu\": 8,\n  \"ram_mb\": 32768,\n  \"gpu_count\": 1,\n  \"gpu_model\": \"A100\",\n  \"plan_id\": \"string\"\n}")
  .asString();
```

```php
<?php
require_once('vendor/autoload.php');

$client = new \GuzzleHttp\Client();

$response = $client->request('POST', 'https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995', [
  'body' => '{
  "name": "ml-training-01",
  "os_distro": "ubuntu",
  "os_type": "linux",
  "template_id": "tmpl_ubuntu_2204_cuda",
  "cpu": 8,
  "ram_mb": 32768,
  "gpu_count": 1,
  "gpu_model": "A100",
  "plan_id": "string"
}',
  'headers' => [
    'Authorization' => 'Bearer <token>',
    'Content-Type' => 'application/json',
    'X-Idempotency-Key' => 'X-Idempotency-Key',
  ],
]);

echo $response->getBody();
```

```csharp
using RestSharp;

var client = new RestClient("https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995");
var request = new RestRequest(Method.POST);
request.AddHeader("X-Idempotency-Key", "X-Idempotency-Key");
request.AddHeader("Authorization", "Bearer <token>");
request.AddHeader("Content-Type", "application/json");
request.AddParameter("application/json", "{\n  \"name\": \"ml-training-01\",\n  \"os_distro\": \"ubuntu\",\n  \"os_type\": \"linux\",\n  \"template_id\": \"tmpl_ubuntu_2204_cuda\",\n  \"cpu\": 8,\n  \"ram_mb\": 32768,\n  \"gpu_count\": 1,\n  \"gpu_model\": \"A100\",\n  \"plan_id\": \"string\"\n}", ParameterType.RequestBody);
IRestResponse response = client.Execute(request);
```

```swift
import Foundation

let headers = [
  "X-Idempotency-Key": "X-Idempotency-Key",
  "Authorization": "Bearer <token>",
  "Content-Type": "application/json"
]
let parameters = [
  "name": "ml-training-01",
  "os_distro": "ubuntu",
  "os_type": "linux",
  "template_id": "tmpl_ubuntu_2204_cuda",
  "cpu": 8,
  "ram_mb": 32768,
  "gpu_count": 1,
  "gpu_model": "A100",
  "plan_id": "string"
] as [String : Any]

let postData = JSONSerialization.data(withJSONObject: parameters, options: [])

let request = NSMutableURLRequest(url: NSURL(string: "https://api.ibee.co.in/v1/compute/gpu-vms?workspace_id=710995")! as URL,
                                        cachePolicy: .useProtocolCachePolicy,
                                    timeoutInterval: 10.0)
request.httpMethod = "POST"
request.allHTTPHeaderFields = headers
request.httpBody = postData as Data

let session = URLSession.shared
let dataTask = session.dataTask(with: request as URLRequest, completionHandler: { (data, response, error) -> Void in
  if (error != nil) {
    print(error as Any)
  } else {
    let httpResponse = response as? HTTPURLResponse
    print(httpResponse)
  }
})

dataTask.resume()
```