AI Provider Integrations Overview

Arize integrates with LLMs from a range of providers to power the following:

Prompt playgrounds
Evaluations
Alyx (Alyx runs on Arize provided models, or on your AI provider)

Note: By adding integrations, your data may be sent to your provider for certain actions within Arize AX (e.g., prompt playground) and your account may be billed for usage.

Supported Providers

Anthropic

AWS Bedrock

Azure OpenAI Service

Doubleword

Gemini

NVIDIA

OpenAI

OpenRouter

OrcaRouter

TrueFoundry

VertexAI

Custom Model Endpoints

Set up your AI Provider Integration

You can set up your AI provider integration using our skills, code, or in Arize AX

Set up an Integration Using Arize Skills

Install the Arize Skills for your coding agent.
Ask your agent to create the relevant integration:
Create an OpenAI integration named 'my-openai' with my API key sk-proj...

Set up an Integration in Arize AX

To configure integrations in Arize AX, Navigate to Settings -> AI Providers.

From here you can add a new integration by selecting your provider. You can also view and edit existing integrations.

Set up an Integration Using Code

CLI
TypeScript SDK
Python SDK
REST

Install the CLI:
pip install arize-ax-cli
Set up your profile (required before running other commands):
ax profiles create
List all AI integrations:
ax ai-integrations list

Create an OpenAI integration

ax ai-integrations create \
    --name "My OpenAI Integration" \
    --provider openAI \
    --api-key $OPENAI_API_KEY

Create an Anthropic integration

ax ai-integrations create \
    --name "My Anthropic Integration" \
    --provider anthropic \
    --api-key $ANTHROPIC_API_KEY

Create an Azure OpenAI integration

ax ai-integrations create \
    --name "My Azure OpenAI Integration" \
    --provider azureOpenAI \
    --api-key $AZURE_OPENAI_API_KEY \
    --base-url "https://my-resource.openai.azure.com/"

Create an AWS Bedrock integration

ax ai-integrations create \
    --name "My Bedrock Integration" \
    --provider awsBedrock \
    --role-arn "arn:aws:iam::123456789012:role/ArizeBedrockRole"

Create a Vertex AI integration

ax ai-integrations create \
    --name "My Vertex AI Integration" \
    --provider vertexAI \
    --project-id "my-gcp-project" \
    --location "us-central1"

Create a Gemini integration

ax ai-integrations create \
    --name "My Gemini Integration" \
    --provider gemini \
    --api-key $GEMINI_API_KEY

Create an NVIDIA NIM integration

ax ai-integrations create \
    --name "My NVIDIA NIM Integration" \
    --provider nvidiaNim \
    --api-key $NVIDIA_API_KEY \
    --base-url "https://integrate.api.nvidia.com/v1"

Create a Custom (OpenAI-compatible) integration

ax ai-integrations create \
    --name "My Custom Integration" \
    --provider custom \
    --base-url "https://my-llm-proxy.example.com/v1" \
    --api-key $CUSTOM_LLM_API_KEY

Install dependencies:
npm install @arizeai/ax-client

Set up the Arize client:

import { createClient } from "@arizeai/ax-client";

const apiKey = process.env.ARIZE_API_KEY;

const client = createClient({
    baseUrl: "https://api.arize.com/v2",
    apiKey: apiKey,
});

List existing integrations:

// List all AI integrations
const integrations = await client.aiIntegrations.list();

Create an integration:

import { createAiIntegration } from "@arizeai/ax-client";

const integration = await createAiIntegration({
    name: "My OpenAI Integration",
    provider: "openAI",
    apiKey: "sk-...",
    modelNames: ["gpt-5.5", "gpt-5.4-mini"],
    enableDefaultModels: true,
});

Install the Arize Python SDK:
pip install 'arize'

Set up the Arize client:

import os
from arize import ArizeClient

api_key = os.environ["ARIZE_API_KEY"]

client = ArizeClient(api_key=api_key)

List all AI integrations

integrations = client.ai_integrations.list()

Create AI integrations

# Create an OpenAI integration
integration = client.ai_integrations.create(
    name="My OpenAI Integration",
    provider="openAI",
    api_key=os.environ["OPENAI_API_KEY"],
)

# Create an Anthropic integration
integration = client.ai_integrations.create(
    name="My Anthropic Integration",
    provider="anthropic",
    api_key=os.environ["ANTHROPIC_API_KEY"],
)

# Create an AWS Bedrock integration
integration = client.ai_integrations.create(
    name="My Bedrock Integration",
    provider="awsBedrock",
    provider_metadata={"role_arn": "arn:aws:iam::123456789012:role/ArizeBedrockRole"},
)

List existing integrations:

curl -X GET "https://api.arize.com/v2/ai-integrations" \
    -H "Authorization: Bearer $ARIZE_API_KEY" \
    -H "Content-Type: application/json"

Create an integration:

curl -X POST "https://api.arize.com/v2/ai-integrations" \
    -H "Authorization: Bearer $ARIZE_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
        "name": "My OpenAI Integration",
        "provider": "openAI",
        "api_key": "'"$OPENAI_API_KEY"'"
    }'

Quickstart

Instrument

Observe

Evaluate

Improve

Agents

Settings

Security

AI Provider Integrations Overview

Supported Providers

Anthropic

AWS Bedrock

Azure OpenAI Service

Doubleword

Gemini

NVIDIA

OpenAI

OpenRouter

OrcaRouter

TrueFoundry

VertexAI

Custom Model Endpoints

Set up your AI Provider Integration

Set up an Integration Using Arize Skills

Set up an Integration in Arize AX

Set up an Integration Using Code

​Supported Providers

Anthropic

AWS Bedrock

Azure OpenAI Service

Doubleword

Gemini

NVIDIA

OpenAI

OpenRouter

OrcaRouter

TrueFoundry

VertexAI

Custom Model Endpoints

​Set up your AI Provider Integration

​Set up an Integration Using Arize Skills

​Set up an Integration in Arize AX

​Set up an Integration Using Code

Supported Providers

Set up your AI Provider Integration

Set up an Integration Using Arize Skills

Set up an Integration in Arize AX

Set up an Integration Using Code