# Create Endpoint - Documentation

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown is available with `Accept: text/markdown` and `.md` URL variants.

Source: /docs/api-reference/serverless-create-endpoint

# Create Endpoint

POST

/

gpu-instance

/

openapi

/

v1

/

endpoint

/

create

Try it

Create Endpoint

cURL

```
curl --request POST \
--url https://api.novita.ai/gpu-instance/openapi/v1/endpoint/create \
--header 'Authorization: <authorization>' \
--header 'Content-Type: <content-type>' \
--data '
{
"endpoint": {
"name": "<string>",
"appName": "<string>",
"workerConfig": {
"minNum": 123,
"maxNum": 123,
"freeTimeout": 123,
"maxConcurrent": 123,
"gpuNum": 123,
"requestTimeout": 123
},
"ports": [
{
"port": "<string>"
}
],
"policy": {
"type": "<string>",
"value": 123
},
"image": {
"image": "<string>",
"authId": "<string>",
"command": "<string>"
},
"products": [
{
"id": "<string>"
}
],
"rootfsSize": 123,
"volumeMounts": [
{
"type": "<string>",
"size": 123,
"id": "<string>",
"mountPath": "<string>"
}
],
"clusterID": "<string>",
"envs": [
{
"key": "<string>",
"value": "<string>"
}
],
"healthy": {
"path": "<string>"
}
}
}
'
```

200

```
{
"id": "<string>"
}
```

##

[​](#request-headers)

Request Headers

[​](#param-content-type)

Content-Type

string

required

Enum: `application/json`

[​](#param-authorization)

Authorization

string

required

Bearer authentication format, for example: Bearer {{API Key}}.

##

[​](#request-body)

Request Body

[​](#param-endpoint)

endpoint

object

required

Endpoint configuration.

Hide properties

[​](#param-name)

name

string

Endpoint name. String with a length limit of 0-220 characters.

[​](#param-app-name)

appName

string

Application name (reflected in the URL). The application name is part of the Endpoint URL, supports customization, and defaults to the Endpoint ID.

[​](#param-worker-config)

workerConfig

object

required

Worker configuration. The valid range is dynamically retrieved through the parameter limits API.

Show properties

[​](#param-min-num)

minNum

integer

required

Minimum number of workers.

[​](#param-max-num)

maxNum

integer

required

Maximum number of workers.

[​](#param-free-timeout)

freeTimeout

integer

required

Idle timeout in seconds.

[​](#param-max-concurrent)

maxConcurrent

integer

required

Maximum concurrency.

[​](#param-gpu-num)

gpuNum

integer

required

Number of GPUs per worker.

[​](#param-request-timeout)

requestTimeout

integer

required

Request timeout (seconds).

[​](#param-ports)

ports

object[]

required

HTTP ports. Only one port is supported. Supported port range: 1-65535, excluding internal ports 2222, 2223, and 2224.

Show properties

[​](#param-port)

port

string

required

HTTP port.

[​](#param-policy)

policy

object

required

Scaling policy. The valid range is dynamically retrieved through the parameter limits API.

Show properties

[​](#param-type)

type

string

required

Scaling policy type. Available values:

- `queue`: Queue latency policy, scales workers based on request wait time in the queue.

- `concurrency`: Queue request policy, scales workers based on the number of requests in the queue.

[​](#param-value)

value

integer

required

The meaning of value depends on the type:

- When type = queue, value represents the queue wait time in seconds.

- When type = concurrency, value represents the maximum number of requests in the queue.

[​](#param-image)

image

object

required

Container image configuration.

Show properties

[​](#param-image-1)

image

string

required

Image address. String with a length limit of 0-511 characters.

[​](#param-auth-id)

authId

string

Private image credential ID (not required for public images or platform user images). String with a length limit of 0-255 characters.

[​](#param-command)

command

string

Container startup command. String with a length limit of 0-2047 characters.

[​](#param-products)

products

object[]

required

Product information.

Show properties

[​](#param-id)

id

string

required

Product ID.

[​](#param-rootfs-size)

rootfsSize

integer

required

Root filesystem size in GB. Currently fixed at 100.

[​](#param-volume-mounts)

volumeMounts

object[]

required

Storage configuration in GB.

Show properties

[​](#param-type-1)

type

string

required

Storage type. Available values:

- `local`: Local storage.

- `network`: Network storage.

[​](#param-size)

size

integer

Local storage size, currently fixed at 30. Not required for network storage.

[​](#param-id-1)

id

string

Network storage ID. Not required for local storage.

[​](#param-mount-path)

mountPath

string

required

Storage mount path. String with a length limit of 0-255 characters.

[​](#param-cluster-id)

clusterID

string

Cluster information. Required when mounting cloud storage and must match the cluster ID where the cloud storage is located. String with a length limit of 0-255 characters.

[​](#param-envs)

envs

object[]

Environment variables.

Show properties

[​](#param-key)

key

string

required

Environment variable name.

[​](#param-value-1)

value

string

required

Environment variable value.

[​](#param-healthy)

healthy

object

required

Health check endpoint configuration.

Show properties

[​](#param-path)

path

string

required

The path to check when performing HTTP health checks.

##

[​](#response)

Response

[​](#param-id-2)

id

string

The created Endpoint ID.

Last modified on March 26, 2026