Run Claude Code with any foundation model on Amazon Bedrock (43 models) or a self-hosted model on EC2. Includes a HumanEval benchmark. - View it on GitHub