Kiro

Kiro is an agentic AI development environment, built and operated by AWS, that moves beyond prompt-based code generation toward what it calls agentic engineering. Instead of jumping straight from a prompt to code, Kiro helps you define requirements, system design, and tasks before implementation—so the code you ship matches what you actually specified.

Key Features

• Spec-driven development: Kiro turns prompts into requirements, architectural designs, and sequenced tasks, then implements them with parallel agents. This explicitly documents reasoning and implementation decisions for your team.
• Correctness checks: Before writing code, Kiro checks requirements for contradictions and gaps using automated reasoning techniques, then validates behavior with property-based tests that assert rules holding across all inputs rather than checking a few examples.
• Parallel agents: Implement tasks with agents on your local machine or in the cloud, with sessions that continue when you step away.
• Model flexibility: Choose from Anthropic Claude, OpenAI GPT, and open-weight models like DeepSeek and Qwen, or use Auto, which picks a model per task based on complexity, quality, latency, and cost. Reasoning effort is adjustable per model.
• Multiple surfaces: Kiro runs as an IDE, a headless CLI for CI/CD workflows (PR reviews and bug fixes without an editor), a web interface with cloud sandboxes, a mobile app, and the open-source Kiro Crew workspace.
• Open standards and extensibility: The agent supports Agent Client Protocol (ACP), AGENTS.md, Skills.md, and MCP. The IDE supports Open VSX extensions, themes, and VS Code settings, and is based on Code OSS so you can import existing VS Code configuration.

Typical Use Cases

• Building a working prototype from a concept: A developer writes a spec, and Kiro generates user stories, design, and tasks before implementing—users report going from concept to prototype in a weekend.
• Large-codebase feature work: Parallel agents and repo-scale models help implement multi-file, multi-agent tasks with documented decisions.
• Automated pipeline tasks: Run the headless CLI in CI/CD to review pull requests and fix bugs without opening an editor.
• Background task automation: Create hooks so tasks like adding unit tests or updating documentation run automatically in the background.

How It Works

1. Describe what you want to build in a prompt.
2. Kiro turns it into requirements, an architectural design, and sequenced tasks, checking requirements for contradictions and gaps.
3. Review the spec, then let parallel agents implement it locally or in the cloud.
4. Validate behavior with property-based tests, then take the work through to a pull request.

Who It's For

Kiro suits developers and teams working on complex or large-codebase projects who want structure and traceability around AI-generated code, as well as engineers automating review and testing in CI/CD pipelines. It supports languages including Python, Java, JavaScript, TypeScript, C#, Go, Rust, PHP, Ruby, Kotlin, C, C++, SQL, and HCL.

Kiro combines spec-driven planning with parallel agents and correctness testing to help teams ship maintainable code that matches their intent.