Source: aws.amazon.com/blogs/machine-learning/use-kimi-code-cli-with-models-on-amazon-bedrock
Kimi Code CLI is an open source coding agent from Moonshot AI. Previously, connecting it to Amazon Bedrock required a generic provider configured with a Bedrock API key, which was a separate credential to obtain and protect. With v2.7.0, the CLI includes a native bedrock provider that can use the standard AWS credential chain, including named profiles and single sign-on.
The provider resolves temporary credentials through the AWS CLI credential chain and generates a short-lived Bedrock bearer token, renewed when needed. Those ephemeral keys are valid for up to 12 hours or the AWS session length, whichever is shorter. The AWS_REGION or AWS_DEFAULT_REGION setting must be set explicitly in the Kimi Code config or shell, since the provider cannot infer it from AWS CLI configuration.
The walkthrough uses the global.moonshotai.kimi-k3.1 inference profile and a config.toml entry with the bedrock provider. The post also covers an IAM policy for the required actions, and notes that the Kimi Code desktop and web apps can show a connected status that does not verify credentials, permissions or model access.
Why it matters if you build agents
Agent developers running Kimi Code on Bedrock can drop the separately managed Bedrock API key and rely on SSO or named AWS profiles, which fits existing AWS security workflows.
Developers should declare model capabilities explicitly in config.toml, since Kimi Code does not auto-detect most Bedrock model capabilities, and should test with a non-interactive prompt rather than trusting the connected status.
Get stories like this every weekday
One short email with the day's agent tooling news.