ZICQ
中 Log in / Sign up
Newsroom Industry & Trends #AWS #Claude Code #Data presence #Amazon Bedrock #IAM

AWS Publishes Guide for Single-Region Claude Code Deployment to Meet Data Residency Compliance

Avatar of Mr.Xu

By Mr.Xu Compiled & Reviewed by Editorial

Published: · 10 views

中文阅读 (Chinese) English Version
Co-reading Select any sentence in the article to highlight, annotate, or copy a quote card. Join discussion ↓

Summary:The AWS Machine Learning Blog has published a technical guide detailing how to configure Claude Code on Amazon Bedrock to ensure all inference requests are processed within a single AWS Region (e.g., London) to meet strict data residency compliance. The article compares two implementation paths: using the Mantle endpoint (available in seven regions including Ireland) and the classic Bedrock Invoke API with application inference profiles (for London), and provides IAM Region condition policies an


Overview

The AWS Machine Learning Blog has published an in-depth technical guide for enterprises with strict data residency compliance requirements, explaining how to configure Claude Code on Amazon Bedrock to ensure all model inference is processed within a specified AWS Region. The article uses a US-headquartered global organization as an example, whose compliance team required all prompts, completions, and intermediate processing to stay within the London (eu-west-2) Region, not merely be initiated from London.

Two Implementation Paths

Path 1: Mantle Endpoint

Mantle is a newer Amazon Bedrock endpoint that provides Anthropic's native API shape. By setting CLAUDE_CODE_USE_MANTLE=1 and AWS_REGION, you can achieve single-Region routing without creating inference profiles. Currently, Mantle supports seven Regions: Ireland, Stockholm, Tokyo, Melbourne, US East (N. Virginia), US East (Ohio), and US West (Oregon), and offers newer models such as Claude Sonnet 5, Opus 4.8, and Haiku 4.5.

Path 2: Classic Bedrock + Application Inference Profile

For Regions not covered by Mantle (e.g., London), you must use the classic Bedrock Invoke API. Since eu-west-2 only supports in-Region inference for Claude Opus 4.6 and Sonnet 4.6, and these models cannot be invoked directly by model ID, you must create application inference profiles. By creating profiles pointing to the London foundation model ARNs via the AWS CLI and setting environment variables, you can pin Claude Code to the London Region.

IAM Region Condition Policy

Both paths require attaching an IAM policy that uses the aws:RequestedRegion condition key to restrict requests to the target Region. The article provides detailed JSON policy examples, ensuring that even if a developer misconfigures the Region, IAM will deny the call.

Compliance Verification

AWS CloudTrail can be used to verify that all calls are made in the target Region. The classic Bedrock path logs InvokeModel events, while the Mantle path logs CreateInference events. The article provides query commands and SQL examples, and recommends enabling data events in CloudTrail Lake.

Choosing the Right Path

The article emphasizes that for most workloads, cross-Region inference (CRIS) is the better default, and single-Region patterns should only be used when compliance requires a specific Region. The choice depends on the target Region: if it is one of Mantle's seven supported Regions, prefer Mantle; if it is London, use classic Bedrock with application inference profiles.

Conclusion

This guide provides a clear and actionable technical solution for enterprises needing data residency compliance, demonstrating how to use native AWS services to enforce single-Region constraints while maintaining full Claude Code functionality.


Source: AWS Machine Learning Blog - Enforcing data residency with single-Region Claude Code on Amazon Bedrock

— END —

Tags: #AWS #Claude Code #Data presence #Amazon Bedrock #IAM

Editorial & Fact-Checking Note: This article is compiled from primary research, official release documentation, and source papers by the ZICQ Newsroom pipeline with automated entity verification and human editorial review. If you notice any technical inaccuracy, please submit a correction via our corrections policy or email our editorial desk directly.

What's your take on this story?
🚀 Explore & Submit Works

Community Comments

Loading live comments and annotations…