Create an AI integration

curl --request POST \
  --url https://api.arize.com/v2/ai-integrations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "name": "Production OpenAI",
  "provider": "OPEN_AI",
  "api_key": "sk-abc123...",
  "model_names": [
    "gpt-4",
    "gpt-4o"
  ],
  "enable_default_models": true,
  "scopings": [
    {
      "organization_id": "QWNjb3VudE9yZzoxMjM6YWJj",
      "space_id": null
    }
  ]
}
'

import requests

url = "https://api.arize.com/v2/ai-integrations"

payload = {
    "name": "Production OpenAI",
    "provider": "OPEN_AI",
    "api_key": "sk-abc123...",
    "model_names": ["gpt-4", "gpt-4o"],
    "enable_default_models": True,
    "scopings": [
        {
            "organization_id": "QWNjb3VudE9yZzoxMjM6YWJj",
            "space_id": None
        }
    ]
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

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

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    name: 'Production OpenAI',
    provider: 'OPEN_AI',
    api_key: 'sk-abc123...',
    model_names: ['gpt-4', 'gpt-4o'],
    enable_default_models: true,
    scopings: [{organization_id: 'QWNjb3VudE9yZzoxMjM6YWJj', space_id: null}]
  })
};

fetch('https://api.arize.com/v2/ai-integrations', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.arize.com/v2/ai-integrations",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'name' => 'Production OpenAI',
    'provider' => 'OPEN_AI',
    'api_key' => 'sk-abc123...',
    'model_names' => [
        'gpt-4',
        'gpt-4o'
    ],
    'enable_default_models' => true,
    'scopings' => [
        [
                'organization_id' => 'QWNjb3VudE9yZzoxMjM6YWJj',
                'space_id' => null
        ]
    ]
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

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

func main() {

	url := "https://api.arize.com/v2/ai-integrations"

	payload := strings.NewReader("{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}")

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

	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(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.arize.com/v2/ai-integrations")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.arize.com/v2/ai-integrations")

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

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}"

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

{
  "id": "TGxtSW50ZWdyYXRpb246MTI6YUJjRA==",
  "name": "My OpenAI Integration",
  "provider": "OPEN_AI",
  "has_api_key": true,
  "model_names": [
    "gpt-4",
    "gpt-4o"
  ],
  "enable_default_models": true,
  "function_calling_enabled": true,
  "auth_type": "DEFAULT",
  "scopings": [
    {}
  ],
  "created_at": "2026-02-13T21:27:19.055Z",
  "updated_at": "2026-02-13T21:27:19.279Z",
  "created_by_user_id": "VXNlcjoxOm5OYkM="
}

{
  "status": 400,
  "title": "Invalid request parameters",
  "detail": "The 'name' field is required and must be a non-empty string.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#invalid-request"
}

{
  "status": 401,
  "title": "Authentication required",
  "detail": "You must be authenticated to access this resource.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#authentication-required"
}

{
  "status": 403,
  "title": "Access forbidden",
  "detail": "You do not have permission to access this resource.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#access-forbidden"
}

{
  "status": 404,
  "title": "Resource not found",
  "detail": "The requested resource with ID '12345' was not found.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#resource-not-found"
}

{
  "status": 409,
  "title": "Resource conflict",
  "detail": "A resource with the given identifier already exists.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#resource-conflict"
}

{
  "status": 422,
  "title": "Unprocessable Entity",
  "detail": "One or more fields failed validation.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#unprocessable-entity"
}

{
  "status": 429,
  "title": "Rate limit exceeded",
  "detail": "You have exceeded the allowed number of requests. Please try again later.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#rate-limit-exceeded"
}

AI Integrations

Create an AI integration

Create a new AI integration for an external LLM provider.

Payload Requirements

name and provider are required.
The integration name must be unique within the account.
provider must be one of: OPEN_AI, AZURE_OPEN_AI, AWS_BEDROCK, VERTEX_AI, ANTHROPIC, NVIDIA_NIM, GEMINI, CUSTOM, LITELLM, FIREWORKS, TOGETHER_AI, DEEP_SEEK, TYPE_SAFE_AI, PERPLEXITY, CLOUDFLARE.
If scopings is omitted, the integration defaults to account-wide visibility.
enable_default_models defaults to false if not provided.
function_calling_enabled defaults to true if not provided.
auth_type defaults to DEFAULT if not provided.
For AWS_BEDROCK provider, provider_metadata must include role_arn.
For VERTEX_AI provider, provider_metadata must include project_id, location, and project_access_label.
TYPE_SAFE_AI, PERPLEXITY, and CLOUDFLARE integrations are used only by Decision Judge evaluators. They require api_key. A base_url, a non-empty model_names or headers, enable_default_models: true, function_calling_enabled: true, or an auth_type other than DEFAULT returns 422. Creating one returns 400 when this deployment has not enabled the provider. That is a deployment setting, separate from the account’s Decision Judge access that the evaluator and task endpoints report as 403.
For CLOUDFLARE provider, provider_metadata must be {"kind": "CLOUDFLARE", "account_id": "<32-character account ID>"}.

Valid example

{
  "name": "Production OpenAI",
  "provider": "OPEN_AI",
  "api_key": "sk-abc123...",
  "model_names": ["gpt-4", "gpt-4o"],
  "enable_default_models": true
}

Invalid example (missing required provider)

{
  "name": "My Integration"
}

This endpoint is in alpha, read more here.

POST

ai-integrations

Create an AI integration

curl --request POST \
  --url https://api.arize.com/v2/ai-integrations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "name": "Production OpenAI",
  "provider": "OPEN_AI",
  "api_key": "sk-abc123...",
  "model_names": [
    "gpt-4",
    "gpt-4o"
  ],
  "enable_default_models": true,
  "scopings": [
    {
      "organization_id": "QWNjb3VudE9yZzoxMjM6YWJj",
      "space_id": null
    }
  ]
}
'

import requests

url = "https://api.arize.com/v2/ai-integrations"

payload = {
    "name": "Production OpenAI",
    "provider": "OPEN_AI",
    "api_key": "sk-abc123...",
    "model_names": ["gpt-4", "gpt-4o"],
    "enable_default_models": True,
    "scopings": [
        {
            "organization_id": "QWNjb3VudE9yZzoxMjM6YWJj",
            "space_id": None
        }
    ]
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

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

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    name: 'Production OpenAI',
    provider: 'OPEN_AI',
    api_key: 'sk-abc123...',
    model_names: ['gpt-4', 'gpt-4o'],
    enable_default_models: true,
    scopings: [{organization_id: 'QWNjb3VudE9yZzoxMjM6YWJj', space_id: null}]
  })
};

fetch('https://api.arize.com/v2/ai-integrations', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.arize.com/v2/ai-integrations",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'name' => 'Production OpenAI',
    'provider' => 'OPEN_AI',
    'api_key' => 'sk-abc123...',
    'model_names' => [
        'gpt-4',
        'gpt-4o'
    ],
    'enable_default_models' => true,
    'scopings' => [
        [
                'organization_id' => 'QWNjb3VudE9yZzoxMjM6YWJj',
                'space_id' => null
        ]
    ]
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

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

func main() {

	url := "https://api.arize.com/v2/ai-integrations"

	payload := strings.NewReader("{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}")

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

	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(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.arize.com/v2/ai-integrations")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.arize.com/v2/ai-integrations")

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

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"name\": \"Production OpenAI\",\n  \"provider\": \"OPEN_AI\",\n  \"api_key\": \"sk-abc123...\",\n  \"model_names\": [\n    \"gpt-4\",\n    \"gpt-4o\"\n  ],\n  \"enable_default_models\": true,\n  \"scopings\": [\n    {\n      \"organization_id\": \"QWNjb3VudE9yZzoxMjM6YWJj\",\n      \"space_id\": null\n    }\n  ]\n}"

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

{
  "id": "TGxtSW50ZWdyYXRpb246MTI6YUJjRA==",
  "name": "My OpenAI Integration",
  "provider": "OPEN_AI",
  "has_api_key": true,
  "model_names": [
    "gpt-4",
    "gpt-4o"
  ],
  "enable_default_models": true,
  "function_calling_enabled": true,
  "auth_type": "DEFAULT",
  "scopings": [
    {}
  ],
  "created_at": "2026-02-13T21:27:19.055Z",
  "updated_at": "2026-02-13T21:27:19.279Z",
  "created_by_user_id": "VXNlcjoxOm5OYkM="
}

{
  "status": 400,
  "title": "Invalid request parameters",
  "detail": "The 'name' field is required and must be a non-empty string.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#invalid-request"
}

{
  "status": 401,
  "title": "Authentication required",
  "detail": "You must be authenticated to access this resource.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#authentication-required"
}

{
  "status": 403,
  "title": "Access forbidden",
  "detail": "You do not have permission to access this resource.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#access-forbidden"
}

{
  "status": 404,
  "title": "Resource not found",
  "detail": "The requested resource with ID '12345' was not found.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#resource-not-found"
}

{
  "status": 409,
  "title": "Resource conflict",
  "detail": "A resource with the given identifier already exists.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#resource-conflict"
}

{
  "status": 422,
  "title": "Unprocessable Entity",
  "detail": "One or more fields failed validation.",
  "instance": "/resource/12345",
  "type": "https://arize.com/docs/ax/rest-reference/errors#unprocessable-entity"
}

{
  "status": 429,
  "title": "Rate limit exceeded",
  "detail": "You have exceeded the allowed number of requests. Please try again later.",
  "instance": "/resource",
  "type": "https://arize.com/docs/ax/rest-reference/errors#rate-limit-exceeded"
}

Authorizations

Authorization

string

header

required

Most Arize AI endpoints require authentication. For those endpoints that require authentication, include your API key in the request header using the format

Service accounts may instead present an access token issued by an external identity provider (Microsoft Entra ID) when the account admin has configured that provider. Send it in the same header: Authorization: Bearer <access-token>.

Body

application/json

Body containing AI integration creation parameters

name

string

required

Integration name

provider

enum<string>

required

The AI provider for this integration. TYPE_SAFE_AI, PERPLEXITY, and CLOUDFLARE integrations are used only by Decision Judge evaluators, and each is available only when that provider is enabled.

Available options:

OPEN_AI,

AZURE_OPEN_AI,

AWS_BEDROCK,

VERTEX_AI,

ANTHROPIC,

CUSTOM,

NVIDIA_NIM,

GEMINI,

LITELLM,

FIREWORKS,

TOGETHER_AI,

DEEP_SEEK,

TYPE_SAFE_AI,

PERPLEXITY,

CLOUDFLARE

api_key

string

API key for the provider (write-only, never returned)

base_url

string

Custom base URL for the provider

model_names

string[]

Supported model names

headers

object

Custom headers to include in requests. The serialized header map must not exceed 8,175 bytes.

Show child attributes

enable_default_models

boolean

Enable provider's default model list (default false)

function_calling_enabled

boolean

Enable function/tool calling (default true)

auth_type

enum<string>

The authentication method for this integration. OAUTH2_CLIENT_CREDENTIALS is returned for integrations configured with OAuth 2.0 client credentials in the Arize UI. Setting it through this API is not yet supported.

Available options:

DEFAULT,

PROXY_WITH_HEADERS,

BEARER_TOKEN,

OAUTH2_CLIENT_CREDENTIALS

provider_metadata

object

AWS Bedrock provider metadata in a write request (strict form of AwsProviderMetadata).

Option 1
Option 2
Option 3

Show child attributes

scopings

object[]

Visibility scoping rules. Defaults to account-wide.

Show child attributes

Response

An AI integration object

An AI integration configures access to an external LLM provider (e.g. OpenAI, Azure OpenAI, AWS Bedrock, Vertex AI). Integrations can be scoped to the entire account, a specific organization, or a specific space.

string

required

The integration ID

name

string

required

The integration name

provider

enum<string>

required

The AI provider for this integration. TYPE_SAFE_AI, PERPLEXITY, and CLOUDFLARE integrations are used only by Decision Judge evaluators, and each is available only when that provider is enabled.

Available options:

OPEN_AI,

AZURE_OPEN_AI,

AWS_BEDROCK,

VERTEX_AI,

ANTHROPIC,

CUSTOM,

NVIDIA_NIM,

GEMINI,

LITELLM,

FIREWORKS,

TOGETHER_AI,

DEEP_SEEK,

TYPE_SAFE_AI,

PERPLEXITY,

CLOUDFLARE

has_api_key

boolean

required

Whether an API key is configured (the key itself is never returned)

enable_default_models

boolean

required

Whether the provider's default model list is enabled

function_calling_enabled

boolean

required

Whether function/tool calling is enabled

auth_type

enum<string>

required

Available options:

DEFAULT,

PROXY_WITH_HEADERS,

BEARER_TOKEN,

OAUTH2_CLIENT_CREDENTIALS

scopings

object[]

required

Visibility scoping rules

Show child attributes

created_at

string<date-time>

required

When the integration was created

updated_at

string<date-time>

required

When the integration was last updated

created_by_user_id

string

required

The user ID of the user who created the integration

base_url

string | null

Custom base URL for the provider

model_names

string[] | null

Supported model names

headers

object | null

Custom headers included in requests

Show child attributes

provider_metadata

object

Provider-specific configuration. For AWS_BEDROCK, must include role_arn. For VERTEX_AI, must include project_id, location, and project_access_label. For CLOUDFLARE, includes account_id.

Option 1
Option 2
Option 3

Show child attributes

Skills

MCP server

API reference

API clients

Create an AI integration

Authorizations

Body

Response