Anthropic Launches Claude Platform on AWS with Managed Agent Tools

Anthropic has officially launched Claude Platform on AWS, marking the first time a cloud provider offers direct access to Anthropic's native platform experience through an existing cloud account. The service reached general availability in May 2026, allowing developers to use Claude's full API suite without managing separate credentials, contracts, or billing relationships.

The announcement comes from AWS's official product update, which details the integration's scope and regional availability. AWS is positioning this as a distinct offering from Claude on Amazon Bedrock, with key architectural differences that matter for enterprise deployment.

Under the new platform, Anthropic operates the service and customer data is processed outside the AWS security boundary. This contrasts with Claude on Amazon Bedrock, where AWS acts as the data processor and keeps all operations within the AWS infrastructure. The distinction becomes critical for organizations with strict regional data residency requirements or those needing data processed exclusively within AWS infrastructure.

Authentication runs through AWS Identity and Access Management, while audit logging is handled through AWS CloudTrail and billing through a single AWS invoice. Developers access the same APIs, features, and console experience available through Anthropic directly, including the Messages API, Claude Managed Agents (beta), advisor tool (beta), web search and web fetch, MCP connector (beta), Agent Skills (beta), code execution, and files API (beta).

The platform includes Claude Managed Agents, a pre-built, configurable agent harness that runs in managed infrastructure. Instead of building your own agent loop, tool execution, and runtime, you get a fully managed environment where Claude can read files, run commands, browse the web, and execute code securely. The harness supports built-in prompt caching, compaction, and other performance optimizations for high-quality, efficient agent outputs.

According to AWS's technical blog post, the setup requires three steps: create a workspace, authenticate, and call the API. A workspace separates projects, environments, or teams while maintaining centralized billing and administration. It also serves as the primary AWS IAM resource for Claude Platform on AWS.

Authentication supports two methods: IAM with AWS Signature Version 4, and API keys. AWS recommends using temporary IAM credentials for setups that require a higher level of security, and API keys for exploring Claude Platform on AWS. After configuration, clients can use capabilities like web search, MCP connectors, agent skills, code execution, and file uploads through Claude Platform on AWS.

Usage monitoring happens through the Claude Console, including breakdowns by workspace, AWS IAM principal, and time period. AWS CloudTrail captures requests to Claude Platform on AWS, whether from the Anthropic SDK, Claude Code, or Cowork. Workspace operations are logged as management events by default, and you can enable data event logging to capture inference activity.

The service 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 (Jakarta), Asia Pacific (Sydney), and Asia Pacific (Melbourne).

Secondary coverage from Digital Watch Observatory notes the launch shows how competition between major AI providers is shifting towards enterprise deployment, cloud integration and agent-based automation. For organisations, the choice is no longer only about model performance, but also about where data is processed, how access is controlled, how audit logs are handled and whether AI agents can be deployed within existing cloud governance systems.

Technical documentation from Claude API Docs clarifies that Claude Managed Agents is currently in beta. All Managed Agents endpoints require the managed-agents-2026-04-01 beta header. The SDK sets the beta header automatically. Behaviors may be refined between releases to improve outputs.

Managed Agents endpoints are rate-limited per organization: 300 requests per minute for create endpoints and 600 requests per minute for read endpoints. Organization-level spend limits and tier-based rate limits also apply. Certain features like outcomes and multiagent are in beta (research preview) and require access requests.

The physical reality of using this platform means developers navigate the AWS Management Console, open the Claude Platform on AWS Console, and create workspaces before making their first API call. Usage is billed through AWS Marketplace on a consumption basis, so you can track and manage AI spending alongside your other AWS services. You can also allocate spending using resource tags.

For teams without specific Regional data residency requirements, Claude Platform on AWS complements Claude models on Amazon Bedrock, so you can access Claude through the approach that fits your needs. The choice between the two depends on whether you prioritize native Anthropic features or strict data boundary compliance.

Whether enterprises actually adopt this over Bedrock remains the real question, especially when compliance teams start asking where their data actually lives.

The launch demonstrates Anthropic's strategy to embed deeper into enterprise cloud workflows while maintaining control over their platform experience. AWS gets a differentiated offering that competes with other cloud-native AI services, and developers get a unified experience without juggling multiple vendor accounts.

Time will tell if the convenience of single-account billing outweighs the data boundary concerns for regulated industries. Until then, developers can test the setup through the AWS Marketplace and send feedback to AWS re:Post or through their usual AWS Support contacts.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn