Tried using the Claude Platform on AWS

The Claude platform on AWS is now generally available. This will allow you to access Claude, Anthropic's native platform, directly through your AWS account. Claude Platform on AWS is now generally available

1. Use Anthropic’s official Claude API directly with your AWS account Claude Platform on AWS is a new offering that allows you to use Anthropic’s native Claude API—operated by Anthropic—seamlessly integrated with AWS IAM authentication, AWS billing, and CloudTrail auditing.
2. Immediate access to the latest features (full functionality, full speed) New features released by Anthropic become available on AWS the same day.
3. Inference runs on Anthropic infrastructure Since Claude Platform on AWS is operated by Anthropic, inference runs outside AWS’s security boundary. Open the “Claude Platform on AWS” page in the AWS Management Console and click “Get Started.” Click “Continue.” Enter your email address and click “Start.” You will receive an email from Anthropic titled “Set up your Claude organization.” Click the link. Enter your organization details and click “Complete setup.” Click “Create Workspace.” Your workspace will be created. Make sure to note the workspace ID. Select the “Admin” role and sign in. Claude Console will open. Let’s take a look at some of the available features. You can register and manage agent skills. You can create and manage agents. You can monitor token usage and costs. You can view rate limits and manage data residency. Generate an API key from the dashboard. Set an expiration date and generate the key. Make sure to save the generated API key. On the “Claude Platform on AWS” page, you can create and manage API keys under “API Keys.” Set your environment variables:

export AWS_REGION="us-east-1"
export ANTHROPIC_WORKSPACE_ID="wrkspc_xxxxx"
export CLAUDE_AWS_BASE_URL="https://aws-external-anthropic.${AWS_REGION}.api.aws"
export ANTHROPIC_API_KEY="xxxxx"
Try asking what Amazon Bedrock is.
curl "https://aws-external-anthropic.us-east-1.api.aws/v1/messages" \
-H "x-api-key: $ANTHROPIC_API_KEY" \
-H "x-amz-security-token: $AWS_SESSION_TOKEN" \
-H "content-type: application/json" \
-H "anthropic-version: 2023-06-01" \
-H "anthropic-workspace-id: $ANTHROPIC_WORKSPACE_ID" \
-d '{
"model": "claude-sonnet-4-6",
"max_tokens": 1024,
"messages": [
{"role": "user", "content": "What is Amazon Bedrock?"}
]

}'

If you receive a response like the one below, it’s working correctly:
{
"model": "claude-sonnet-4-6",
"id": "msg_xxxxx",
"type": "message",
"role": "assistant",
"content": [
{
"type": "text",
"text": "# Amazon Bedrockとは\n\nAmazon Bedrockは、**AWSが提供する完全マネージド型の生成AIサービス**です。\n\n---\n\n## 主な特徴\n\n### 🤖 複数の基盤モデル（Foundation Models）へのアクセス\n様々なAIプロバイダーのモデルを一つのAPIで利 用できます：\n\n| プロバイダー | モデル例 |\n|------------|---------|\n| Amazon | Amazon Titan |\n| Anthropic | Claude 3.5など |\n| Meta | Llama 3など |\n| Mistral AI | Mistral / Mixtralなど |\n| Stability AI | Stable Diffusionなど |\n\n---\n\n## 主な機能\n\n- **テキスト生成** - 文章作成、要約、翻訳など\n- **画像生成** - テキストから画像を生成\n- **RAG（検索拡張生成）** - 独自データと組み合わせた回答生成\n- **AIエージェント** - 複雑なタスクの自動化\n- **Fine-tuning** - 独自データでモデルをカスタマイズ\n\n---\n\n## メリット\n\n✅ **サーバーレス** - インフラ管理不要 \n✅ **セキュリティ** - データはAWS内で保護 \n✅ **スケーラビリティ** - 需要に応じて自動スケール \n✅ **コスト効率** - 使った分だけ支払い\n\n---\n\n## ユースケース例\n\n- チャットボット・カスタマーサポート\n- ドキュメント分析・要約\n- コード生成\n- コンテンツ作成\n\nAWSのエコシステム（S3、Lambda等）と連携しやすいのも大きな利点です。\n\n何か具体的に知りたい点はありますか？"
}
],
"stop_reason": "end_turn",
"stop_sequence": null,
"stop_details": null,
"usage": {
"input_tokens": 18,
"cache_creation_input_tokens": 0,
"cache_read_input_tokens": 0,
"cache_creation": {
"ephemeral_5m_input_tokens": 0,
"ephemeral_1h_input_tokens": 0
},
"output_tokens": 527,
"service_tier": "standard",
"inference_geo": "global"
}
}

Being able to use the latest native Claude features directly with your AWS account is surprisingly smooth, and the setup is incredibly simple. This greatly expands the range of development possibilities. Since inference runs on Anthropic’s side, it’s important to choose between Bedrock and Claude Platform on AWS depending on your use case. It’s a good idea to become comfortable using Claude in various patterns.