These are my personal notes for configuring Claude Code for use with AWS Bedrock. I've also included basic usage and advanced use cases.
I will try to update periodically, but if you notice any errors, please feel free to submit a PR, file an issue, or contact me so I can update.
AWS Bedrock provides a single interface for foundation models from Anthropic, Meta, Mistral, Stability AI, and Amazon, allowing for text, image, and chat-based applications without managing infrastructure. When using Claude Code with AWS Bedrock, Claude runs within our AWS infrastructure, giving us control over data residency and security.
Note
Before proceeding, you should be able to login to AWS console in a browser via SSO and see your account number and a role that has access to Bedrock. Ask your AWS administrator if you're not sure where to find these.
- AWS account number
- Name of AWS Role with Bedrock access
- SSO Start URL
- AWS CLI v2
Follow the steps below to configure your AWS profile, enable SSO authentication for AWS CLI, install Claude Code, and configure Claude to use Bedrock.
Once configured, you can start Claude Code by changing to your project folder and running:
claude
See troubleshooting for more help with common errors when using claude with bedrock.
Reference useful AWS CLI commands.
See basic usage for common commands and pro tips for getting started with Claude Code including:
- Common Commands
- Adding Context With Config Files (claude.md & skills)
- AI-assisted Development Best Practices
See advanced usage for more advanced workflows with claude code including:
