AWS launches Kiro, a Claude-powered AI IDE for developers

Amazon Web Services (AWS) has launched Kiro, a new integrated development environment (IDE) that incorporates artificial intelligence to support structured, spec-driven software development.

The tool is currently available in public preview and is positioned to support developers working on complex codebases with AI-assisted planning and execution.

Focus on structured workflows and planning

Kiro introduces a workflow where developers begin with a high-level prompt, which the system then translates into structured software requirements. These include user stories using the EARS format, technical design documents, and task lists with testing and accessibility considerations.

This spec-first approach is intended to provide more traceable, maintainable, and verifiable software projects, differing from tools that focus only on code generation from natural language.

Integrated automation and agentic support

The IDE includes automation features through “agent hooks,” allowing Kiro to automatically update documentation, generate tests, or perform security checks when files are modified or new commits are made. These automations aim to maintain quality standards without requiring additional manual steps.

Developers can also interact with the AI system using files, documents, or URLs as input, enabling multi-modal queries and task execution. Kiro supports agentic workflows, where AI can plan and carry out multi-step development tasks in response to user intent.

Compatibility and model integration

Kiro is built on Code OSS, the open-source foundation of Microsoft Visual Studio Code, and is compatible with existing extensions and developer settings. It currently uses Anthropic’s Claude Sonnet 3.7 and Claude 4.0 models. AWS has not ruled out support for other model providers in the future.

The tool supports the Model Context Protocol (MCP), allowing integration with external systems. It is currently available for macOS (both Intel and Apple Silicon), Windows, and Linux.

Kiro does not require AWS credentials to begin using the platform. Developers can sign in using social logins such as GitHub or Google accounts.

Pricing model and availability

Kiro is in public preview with an initial free tier offering 50 agent interactions per month. Additional paid tiers include:

• Kiro Pro: USD 19/month for 1,000 interactions
• Kiro Pro+: USD 39/month for 3,000 interactions
• Pay-as-you-go: USD 0.04 per interaction beyond plan limits

The tool is being released alongside AWS’s Q Developer tool, which focuses on code-level assistance within existing IDEs. In contrast, Kiro supports broader agentic workflows from specification to implementation.

Industry context and reception

Kiro enters a growing field of AI-powered development tools, including GitHub Copilot, Cursor, Claude Code, and the recently acquired Windsurf. These tools are part of a larger shift towards agentic AI systems capable of handling more comprehensive software engineering tasks.

The public preview of Kiro is now live, with AWS expected to gather feedback before a wider rollout.