Reactive Machines

Introducing Claude's Platform on AWS: Anthropic's native platform, with your AWS account

Today, we are excited to announce the general availability of the Claude Platform on AWS. Claude Platform on AWS is a new service that gives customers direct access to the Anthropic Claude Platform platform through their AWS account, without separate credentials, contracts, or payment relationships required. AWS is the first cloud provider to offer access to native Claude Platform information.

In this post, we explore how the Claude Platform on AWS works and how you can start using it today.

Claude Platform experience with AWS

With the Claude Platform on AWS, you work with the same APIs, features, and console experience available through Anthropic directly. This includes Messages API, Claude Managed Agents (beta), a counselor a tool (beta), web search again web download, MCP connector (beta), Agent skills (beta), code execution, API files (beta). For a full list of skills, see i Claude Platform documents.

You access the Claude Platform on AWS with standard AWS features:

  • Confirmation: You use your existing AWS IAM credentials to access the Claude Platform. There are no separate accounts or API keys to manage.
  • Charging: Usage is billed through AWS Marketplace on a per-use basis, so you can track and manage AI usage alongside your other AWS services.
  • Audit: Activity is captured in AWS CloudTrail, so you can monitor, test, and investigate AI usage the same way you do for other AWS services.

The Claude Platform on AWS is powered by Anthropic, and the underlying applications and data are processed outside the AWS security perimeter. This makes it well-suited for teams that don't have specific data residency requirements for the Region, and it's compatible with Claude's models on Amazon Bedrock, so you can access Claude using a method that suits your needs.

Figure 1: Log in to the AWS Management Console and open a file Claude Platform in the AWS Console

Getting Started with the Claude Platform on AWS

You can run the Claude Platform on AWS through the AWS Marketplace. For step-by-step instructions, see Set up your account. After your account is activated, getting to your first API call takes three steps: create a workspace, authenticate, and call the API.

Step 1: Create a workspace

With a workspace, you can separate projects, areas, or teams while maintaining centralized billing and management. It also serves as the primary AWS Identity and Access Management (IAM) resource for the Claude Platform on AWS. You grant or deny access to specific workspaces through IAM policies using the workspace's ARN. See IAM policies for policy examples.

Open the Claude Console within the Claude Platform in the AWS Console and create a workspace.

Figure 2: You can find the workplace ID at Claude Platform is open AWS Console.

Step 2: Confirm

Claude Platform on AWS supports two authentication methods: IAM with AWS Signature Version 4, and API keys. We recommend using temporary IAM credentials in setups that require a high level of security, and API keys to test the Claude Platform on AWS.

To quickly test your setup, you can create an API key for Claude Platform in the AWS Console:

Figure 3: You can generate an API key for Claude Platform in the AWS Console

Set your API key, base URL, and workspace ID as environment variables:

# Your API key  
export ANTHROPIC_API_KEY= 

# Your regional endpoint for Claude Platform on AWS 
export ANTHROPIC_BASE_URL=https://aws-external-anthropic..api.aws 

# Your workspace ID (find in Claude Platform on AWS Console → Workspaces) 
export ANTHROPIC_WORKSPACE_ID= 

Step 3: Make your first API call

You can now enter i Anthropic Client SDKs again do API calls:

from anthropic import Anthropic 
import os 
 
client = Anthropic( 
   default_headers={"anthropic-workspace-id": os.environ["ANTHROPIC_WORKSPACE_ID"]}, 
) 
 
message = client.messages.create( 
   model="claude-sonnet-4-6", 
   max_tokens=1024, 
   messages=[{"role": "user", "content": "Hello!"}], 
) 
 
print(message) 

Look Starting notebooks for additional code examples.

Claude Platform on AWS in practice

Once your setup is complete, you can point to Claude Code, Claude Cowork, or any other API client in your workspace using the following environment variables or configuration:

export ANTHROPIC_API_KEY= 

export ANTHROPIC_BASE_URL=https://aws-external-anthropic..api.aws 

 # For Claude Cowork, set the “anthropic-workspace-id" in your inference configuration. For Claude Code use the following:  
export ANTHROPIC_CUSTOM_HEADERS='{"anthropic-workspace-id":""}' 

 # For the Anthropic SDK 
export ANTHROPIC_WORKSPACE_ID= 

After you're connected, your customers can use capabilities like web search, MCP connectors, agent capabilities, coding, and file uploads with the Claude Platform on AWS.

Figure 4a: Claude Cowork connected to Claude Platform on AWS

Figure 4b: Claude code connected to the Claude Platform on AWS

You can monitor usage in the Claude Console, including workspace classification, AWS IAM principal, and time.

Figure 5: Usage statistics on Claude Console

In your AWS environment, AWS CloudTrail captures requests from the Claude Platform on AWS, whether they come from the Anthropic SDK, Claude Code, or Cowork. Workspace activities are included as event handlers by default, and you can enable data event logging to capture activity for consideration. For details on event types and logging configuration, see Monitoring and logging. Because usage is billed through AWS Marketplace, you can monitor costs in AWS Cost Explorer alongside your other cloud services. You can also allocate money using resource tags.

Figure 6: Example of AWS CloudTrail (left) and AWS Cost Explorer (right) events for the Claude Platform on AWS

The conclusion

With Claude Platform on AWS, your teams get full native Anthropic APIs and features with the same AWS account you already use. Claude Platform on AWS is available in US East (N. Virginia), US East (Ohio), US West (Oregon), Canada (Central), South America (São Paulo), Europe (Dublin), Europe (London), Europe (Frankfurt), Europe (Milan), Europe (Zurich), Europe (Paris), Europe (Stockholm), Asia Pacific (Tokyo), Asia Pacific (Seoul), Asia Pacific (Seoul), Asia Pacific (Mel Pacific), Asia Pacific (Mel Pacific), Asia Pacific, Asia Pacific (Melbourne). To get started, open the Claude Platform in the AWS Console or check out the documentation.

Try the Claude Platform on AWS today and send feedback to AWS too: Submit or through your regular AWS support contacts.


About the writers

Dani Mitchell

Dani Mitchell is a Sr GenAI Specialist Solutions Architect at AWS and SA leading Amazon Bedrock Knowledge Bases. He helps businesses around the world design and implement productive AI solutions using Amazon Bedrock and Anthropic's models to build scalable, production-ready applications.

Sofian Hamiti

Sofian Hamiti is a technology leader with over 12 years of experience building AI solutions, and leads high-performance teams to maximize customer results. He is passionate about empowering diverse talents to drive global impact and realize their career aspirations.

Eugenio Soltero

Eugenio Soltero is Sr. Product Marketing Manager for Amazon Bedrock at AWS. With several years of experience in generative AI, he helps clients navigate the evolving landscape of foundational models and generative ai to find solutions that deliver measurable value.

Antonio Rodriguez

Antonio Rodriguez is the principal leader of Generative AI Tech at Amazon Web Services. He helps companies of all sizes solve their challenges, embrace innovation, and create new business opportunities with Amazon Bedrock. Outside of work, he likes to spend time with his family and play sports with his friends.

Ayan Ray

Ayan Ray is the Principal Architect of Partner Solutions and AI Tech Lead at AWS, serving as the Global Technical Lead for Anthropic at AWS. He works at the intersection of cloud architecture and Artificial Intelligence, helping organizations adopt and scale Anthropic technologies on AWS.

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button