AWS launches Kiro AI coding assistant with new IDE, CLI tools

AWS has announced the general availability of Kiro, its AI coding assistant, with new features aimed at improving code quality and collaboration for both individual developers and larger teams. The update includes enhancements to the Kiro Integrated Development Environment (IDE), a new Command Line Interface (CLI), and expanded access for organisations and startups.

Code quality focus

The latest version of the Kiro IDE introduces property-based testing to verify if code behaves according to defined specifications. This approach aims to address a common issue with traditional unit tests, which typically validate only specific scenarios and may overlook edge cases. Property-based testing analyses project specifications to extract properties that describe how the system should generally behave and generates a wide range of random test cases based on those properties.

By using this method, Kiro attempts to surface more potential errors and ensure code behaviour aligns with what was intended in the original specification. If violations are found, the system can suggest updates to code implementations or to the specifications themselves. According to AWS, these tests provide a broader evidence of correctness than can be achieved manually, although they do not offer formal verification or proof.

Project management tools

Kiro's new checkpointing feature allows developers to roll back to previous stages of their agent's execution flow. Every significant change or action by the development agent is recorded as a checkpoint. This enables users to undo steps without losing subsequent work or incurring additional credit costs to repeat earlier actions.

Additionally, multi-root workspace support has been added, allowing teams to use Kiro across projects that span multiple roots or submodules. This is intended to benefit organisations with large codebases or multiple packages, who can now work across different components within a single Kiro workspace.

Terminal-based workflow

The new Kiro CLI brings AI-powered development tools directly to the terminal environment, supporting both macOS and Linux systems. Through the CLI, developers can interact with Kiro agents to build features, automate workflows, analyse errors, and trace bugs using familiar command-line tools. The CLI supports custom agents optimised for specific tasks, allowing users to tailor assistants for particular code areas such as frontend or backend development.

Features from the Kiro IDE, including steering files, advanced context management, and Multiple Context Providers (MCP) tools, are available in the CLI. Users can share their Kiro subscriptions and credit allocations across both the IDE and CLI.

Organisational integration

Kiro can now be integrated with AWS Identity Centre, enabling organisations to centrally manage team access and subscriptions. Administrators can assign different Kiro subscription levels, monitor usage and costs, and control account permissions from a single management dashboard. Enhanced identity provider support is planned for future updates.

Startup support

AWS is extending offers to qualifying startups, who can apply for up to a year's worth of Kiro Pro+ credits. This incentive is available globally to startups up to Series B funding rounds. Existing AWS Activate credits can also be applied towards Kiro, and both offers can be combined where applicable.

"Across teams, tools, and tests, Kiro now better supports the way you want to work by bringing the right level of context and structure to AI-powered development. This is just the start," said Nikhil Swaminathan, Product Lead, AWS.