AWS describes Amazon Bedrock cost attribution and usage controls for AI tools

AWS Machine Learning described Amazon Bedrock’s built-in cost attribution controls in a blog post. The system traces each inference call back to the IAM principal, then carries identity and session tags into CUR 2.0 and Cost Explorer without custom instrumentation. AWS said the same approach works for tools such as Claude Code, Claude Desktop, OpenAI Codex, ChatGPT Work, and OpenCode when they call Bedrock through bedrock-runtime or bedrock-mantle. The post also outlines enforcement options using AWS Budgets, CloudWatch, Lambda, and DynamoDB.