What happened
Moonshot AI released version 2.7.0 of its open‑source Kimi Code CLI, introducing a native Amazon Bedrock provider. The new provider lets the CLI use the standard AWS credential chain—including named profiles and single sign‑on—rather than requiring a dedicated Bedrock API key. The blog post walks through configuring the provider, setting IAM permissions, and running a test request against the Kimi K3.1 Pro Highspeed model on Bedrock. The update also clarifies regional settings, cross‑region profiles, and capability declarations needed for proper model use.
Moonshot AI announced Kimi Code CLI v2.7.0, which adds a native Amazon Bedrock provider. Previously, users had to configure a generic provider with an explicit Bedrock API key, adding an extra credential to manage.
The native provider integrates with the standard AWS credential chain, supporting named profiles, SSO, and short‑term credentials. Authentication follows a four‑step process that automatically generates temporary Bedrock tokens valid for up to 12 hours or the session length.
The blog provides a step‑by‑step guide: setting up AWS CLI credentials, creating an IAM policy granting Bedrock permissions for the Kimi K3.1 Pro Highspeed model, configuring the provider in Kimi Code’s config.toml, and sending a test request. It also notes that the provider requires an explicit AWS_REGION setting and does not infer region from the default AWS config.
The update clarifies that while the provider removes the need for a separate API key, users remain responsible for protecting AWS credentials and applying least‑privilege IAM policies. It also highlights that Bedrock is billed per , with global cross‑region profiles offering lower costs.
Source details: aws.amazon.com
Why it matters
The change simplifies the workflow for developers and enterprises already using AWS, reducing credential sprawl and aligning with AWS best practices that favor short‑term, automatically rotated credentials. By leveraging the existing AWS credential chain, organizations can enforce consistent IAM policies, audit access through a single identity provider, and avoid the operational overhead of managing separate Bedrock API keys. This also lowers the barrier for teams to adopt Kimi Code for code‑generation tasks on Bedrock, potentially expanding the ecosystem of AI‑assisted development tools that run within AWS’s security and compliance perimeter. However, users must still configure appropriate IAM permissions and be aware of regional costs, as the provider does not auto‑detect model capabilities and requires explicit capability flags.
Credential management: By using the existing AWS credential chain, organizations can avoid the operational risk of storing and rotating a separate Bedrock API key, aligning with AWS’s recommendation for short‑term credentials.
Security posture: Centralizing authentication through IAM enables unified audit logs, tighter access controls, and easier compliance reporting for regulated workloads that must stay within AWS data boundaries.
Developer experience: The native provider reduces setup friction for developers already familiar with AWS tooling, potentially accelerating adoption of AI‑assisted coding within existing CI/CD pipelines.
Cost awareness: The post emphasizes reviewing Bedrock pricing and choosing global profiles for lower rates, which is critical for teams scaling ‑based usage.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').A route planner searches possible journeys using explicit rules. What does this illustrate about AI?
What to watch next
Future updates may add UI support for the new provider in Kimi Code’s desktop and web apps, reducing the need to edit config files manually. Watch for broader model support beyond Kimi K3.1 Pro, and for any changes to Bedrock pricing or cross‑region policies that could affect cost calculations. Organizations should monitor IAM policy revisions from AWS that could tighten permissions for Bedrock calls, as well as any security advisories related to credential handling in the native provider.
UI integration: Current desktop and web versions of Kimi Code still require manual editing of config.toml for the Bedrock provider. Future releases may expose these settings in the UI, simplifying adoption.
Model expansion: The guide focuses on Kimi K3.1 Pro Highspeed. Monitoring support for additional Bedrock models will indicate how broadly the provider can be used.
Pricing and policy changes: Amazon Bedrock’s pricing and cross‑region policies may evolve, affecting cost calculations for heavy users.
IAM policy updates: AWS may introduce stricter permission scopes for Bedrock, requiring users to adjust their policies to maintain access.