Amazon Web Services has detailed how development teams working on regulated government and compliance-heavy workloads can connect Anthropic's Claude Code tool to Amazon Bedrock inside AWS GovCloud (US) regions. The integration supports Anthropic's Claude Opus 5.5, Claude Sonnet 5.5, and Claude Sonnet 5 models, allowing developers to execute agentic coding tasks while maintaining compliance alignment for sensitive cloud environments.
Endpoint Architecture and Compliance Levels
Amazon Bedrock in AWS GovCloud (US) operates through two distinct API endpoints, both built on the Mantle inference engine featuring Zero Operator Access architecture. The bedrock-runtime endpoint uses standard AWS SDK APIs and provides access to Amazon Bedrock Guardrails, invocation logging, and Knowledge Bases. It is available across both the US-West and US-East GovCloud regions, making it the primary path for organizations requiring compliance audit trails.
Alternatively, the bedrock-mantle endpoint, available in the US-West region, natively implements the Anthropic Messages API. This surface gives development teams access to features such as server-side tools, background inference, and Projects, though it currently omits Bedrock Guardrails and invocation logging. Regarding compliance coverage, Claude Sonnet 5 maintains DoD Impact Level 4 and 5 (IL4/IL5) authorization alongside FedRAMP Class D certification, while Claude Opus 5.5 and Claude Sonnet 5.5 hold FedRAMP Class D certification.
Developer Tooling and Setup Workflows
Claude Code functions directly inside developer terminals, IDEs like VS Code and JetBrains, or continuous integration environments like GitHub Actions and GitLab CI/CD. The agent reads local codebases, executes shell commands, runs test suites, resolves merge conflicts, and interfaces with external tooling using the Model Context Protocol.
Teams can configure Claude Code to route requests through Amazon Bedrock via an interactive terminal setup wizard or by setting system environment variables such as CLAUDE_CODE_USE_BEDROCK and target AWS regions. Developers can pin specific model identifiers within their configuration files to establish consistent model usage across engineering groups.
Security Controls and Cost Guardrails
Because Claude Code executes on local workstations while Bedrock secures model inference, enterprise guidelines recommend using AWS IAM Identity Center to manage single sign-on access with temporary credentials. Administrators can deploy mandatory organization-wide permissions and shared CLAUDE.md memory files across repositories to enforce unified safety and coding standards.
To manage operational costs, AWS notes that unpinned Claude Code configurations default to Claude Opus 5.5, which incurs higher per-token pricing. Enterprise deployments can set Claude Sonnet 5.5 or Claude Sonnet 5 as default models, enable prompt caching, and set up real-time token usage guardrails using Amazon CloudWatch, AWS Lambda, and Amazon DynamoDB.
AI-generated editorial illustration; not a documentary photograph.
Source: Supercharge regulated workloads with Claude Code and Amazon Bedrock
SIGNAL TITANS
Leave a Reply