Create an integration

Create a new integration. The type field selects the config shape; for LLM, config.provider selects the per-provider config. Supported types: LLM, AGENT, EVALUATOR.

Payload Requirements

• type, name, and config are required.
• name must be unique within the account. LLM names are unique among LLM integrations; AGENT and EVALUATOR names share the same remote endpoint integration namespace.
• Server-set fields (id, created_at, updated_at) are not accepted on input.
• scopings defaults to account-wide visibility when omitted.
• For type: EVALUATOR, config.endpoint and config.input_schema are required. config.headers is optional (omit means no headers; pass null or {} on PATCH to clear them). config.headers is encrypted at rest and never returned — reads surface has_headers instead.
• type: EVALUATOR requires the remote evaluators feature to be enabled for the account. If not enabled the request returns 403.

Valid example (type: EVALUATOR)

{
  "type": "EVALUATOR",
  "name": "My remote evaluator",
  "config": {
    "endpoint": "https://eval.example.com/evaluate",
    "input_schema": { "type": "object" }
  }
}

Invalid example (missing required config.endpoint)

{
  "type": "EVALUATOR",
  "name": "My remote evaluator",
  "config": {
    "input_schema": { "type": "object" }
  }
}

Invalid example (missing required config)

{
  "type": "EVALUATOR",
  "name": "My remote evaluator"
}

This endpoint is in alpha, read more here.

POST /v2/integrations

cURL

create_llm

curl --request POST \ --url https://api.arize.com/v2/integrations \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: application/json' \ --data ' { "type": "LLM", "name": "Production OpenAI", "config": { "provider": "OPEN_AI", "api_key": "sk-abc123..." } } '

201

llm_integration

{ "id": "TGxtSW50ZWdyYXRpb246MTI6YUJjRA==", "type": "LLM", "name": "Production OpenAI", "scopings": [ { "organization_id": null, "space_id": null } ], "created_at": "2026-02-13T21:27:19.055Z", "updated_at": "2026-02-13T21:27:19.279Z", "created_by_user_id": "VXNlcjoxOm5OYkM=", "config": { "provider": "OPEN_AI", "has_api_key": true, "is_function_calling_enabled": true } }

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

Body

application/json

Create a new integration. The type field selects the config shape; for LLM, config.provider selects the per-provider config.

Payload Requirements

• type, name, and config are required.
• name must be unique within the account. LLM names are unique among LLM integrations; AGENT and EVALUATOR names share the same remote endpoint integration namespace.
• scopings defaults to account-wide visibility when omitted.
• Type- and provider-specific rules (required fields, defaults, write-only secrets) are documented on each config schema: see the per-provider members of CreateLlmConfig for type: LLM, CreateAgentConfig for type: AGENT, and CreateEvaluatorIntegrationConfigInput for type: EVALUATOR.

Valid example

{
  "type": "LLM",
  "name": "Production OpenAI",
  "config": {
    "provider": "OPEN_AI",
    "api_key": "sk-abc123..."
  }
}

Invalid example (missing required config)

{
  "type": "LLM",
  "name": "Bad Integration"
}

Invalid example (missing required config.provider for type: LLM)

{
  "type": "LLM",
  "name": "Bad Integration",
  "config": {}
}

Option 1

Option 2

Option 3

type · enum<string> · required

Discriminator identifying this request as an LLM integration.

Available options: LLM

name · string · required

Integration name. Unique per (account, type).

config · object · required

Provider-specific configuration for the LLM integration, selected by config.provider.

Option 1

Option 2

Option 3

Option 4

Option 5

Option 6

Option 7

Option 8

Option 9

Show child attributes

scopings · object[]

Visibility scoping rules. Defaults to account-wide.

Show child attributes

Response

201

application/json

An integration object.

Option 1

Option 2

Option 3

A polymorphic integration resource. The type field selects the config shape; for LLM, config.provider selects the per-provider config.

id · string · required

The unique identifier for the integration.

type · enum<string> · required

Discriminator identifying an LLM integration.

Available options: LLM

name · string · required

The integration name. Unique per (account, type).

scopings · object[] · required

Visibility scoping rules. Account-wide when empty.

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

Unique identifier of the user who created the integration.

config · object · required

Per-provider LLM config, discriminated by provider.

Option 1

Option 2

Option 3

Option 4

Option 5

Option 6

Option 7

Option 8

Option 9

Show child attributes