Anthropic SDK

Prerequisite: Complete the Quickstart to get an access key. All requests go to https://gateway.ngrok.ai.

The AI Gateway is compatible with Anthropic’s official SDKs. Set the base URL to https://gateway.ngrok.ai and authenticate with your access key.

Anthropic models also work through OpenAI-compatible SDKs (OpenAI SDK, Vercel AI SDK, LangChain, etc.)—just use the model name directly, for example model: "claude-opus-4-6". You don’t need the Anthropic SDK to access Anthropic models.

Installation

pip install anthropic

npm install @anthropic-ai/sdk

Basic usage

Point the SDK at your AI Gateway endpoint:

import anthropic

client = anthropic.Anthropic(
    base_url="https://gateway.ngrok.ai",
    api_key="ng-xxxxx-g1-xxxxx"  # Your access key
)

message = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)

print(message.content)

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://gateway.ngrok.ai",
  apiKey: "ng-xxxxx-g1-xxxxx",  // Your access key
});

const message = await client.messages.create({
  model: "claude-opus-4-6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello!" }],
});

console.log(message.content);

Try using this prompt on your agent to verify your Anthropic SDK connection: 'State your exact model name and provider.'

Open in Cursor

The Anthropic SDK sends the API key via the x-api-key header. Access keys work with both x-api-key and Authorization: Bearer headers.

Streaming

The AI Gateway supports streaming responses with the Anthropic SDK:

import anthropic

client = anthropic.Anthropic(
    base_url="https://gateway.ngrok.ai",
    api_key="ng-xxxxx-g1-xxxxx"  # Your access key
)

with client.messages.stream(
    model="claude-opus-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a haiku about APIs"}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://gateway.ngrok.ai",
  apiKey: "ng-xxxxx-g1-xxxxx",  // Your access key
});

const stream = await client.messages.create({
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Hello, Claude' }],
  model: 'claude-opus-4-6',
  stream: true,
});
for await (const messageStreamEvent of stream) {
  console.log(messageStreamEvent.type);
}

Try using this prompt on your agent to test streaming — tokens should appear one by one: 'Write a haiku about APIs'

Open in Cursor

Model failover

Specify fallback models in the request body. See Configure fallback models.

When using Anthropic SDKs with the gateway, model selection is limited to providers that support the Anthropic Claude API. Currently in the ngrok model catalog Anthropic is the only provider that supports this format.

Tool use

Tool calling works as documented by Anthropic:

import anthropic

client = anthropic.Anthropic(
    base_url="https://gateway.ngrok.ai",
    api_key="ng-xxxxx-g1-xxxxx"  # Your access key
)

message = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "What's the weather in Paris?"}],
    tools=[{
        "name": "get_weather",
        "description": "Get weather for a location",
        "input_schema": {
            "type": "object",
            "properties": {
                "location": {"type": "string"}
            },
            "required": ["location"]
        }
    }]
)

print(message)

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://gateway.ngrok.ai",
  apiKey: "ng-xxxxx-g1-xxxxx",  // Your access key
});

const message = await client.messages.create({
  model: "claude-opus-4-6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "What's the weather in Paris?" }],
  tools: [
    {
      name: "get_weather",
      description: "Get weather for a location",
      input_schema: {
        type: "object",
        properties: {
          location: { type: "string" },
        },
        required: ["location"],
      },
    },
  ],
});

console.log(message);

Error handling

The gateway handles many errors automatically through failover. For errors that reach your app:

import anthropic

client = anthropic.Anthropic(
    base_url="https://gateway.ngrok.ai",
    api_key="ng-xxxxx-g1-xxxxx"  # Your access key
)

try:
    message = client.messages.create(
        model="claude-opus-4-6",
        max_tokens=1024,
        messages=[{"role": "user", "content": "Hello!"}]
    )
except anthropic.RateLimitError as e:
    print("Rate limited across all providers")
except anthropic.APIError as e:
    print(f"API error: {e}")

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://gateway.ngrok.ai",
  apiKey: "ng-xxxxx-g1-xxxxx",  // Your access key
});

try {
  const message = await client.messages.create({
    model: "claude-opus-4-6",
    max_tokens: 1024,
    messages: [{ role: "user", content: "Hello!" }],
  });
} catch (error) {
  if (error instanceof Anthropic.RateLimitError) {
    console.log("Rate limited across all providers");
  } else if (error instanceof Anthropic.APIError) {
    console.log(`API error: ${error.message}`);
  }
}

Supported endpoints

The AI Gateway supports these Anthropic Claude API endpoints:

Endpoint	Description
`/v1/messages`	Messages

Next steps

Choose a model: Model IDs and provider prefixes
Configure fallback models: Try another model when the first one fails
Choose how to reach providers: Credits, provider keys, and self-run models

Concepts

Getting Started With

SDKs

Guides

Observability

Examples

Reference

API Reference

Installation

Basic usage

Streaming

Model failover

Tool use

Error handling

Supported endpoints

Next steps

​Installation

​Basic usage

​Streaming

​Model failover

​Tool use

​Error handling

​Supported endpoints

​Next steps

Installation

Basic usage

Streaming

Model failover

Tool use

Error handling

Supported endpoints

Next steps