Kiro is an agentic engineering platform and CLI coding agent built and operated by AWS. It extends beyond standard code-completion assistants by implementing a spec-driven development workflow, translating user prompts into structured requirements, architectural designs, and sequenced execution tasks. In daily operations, developers interact with Kiro via its standalone CLI, the Kiro Crew workspace, or dedicated IDE integration. The agent orchestrates subagents to implement tasks locally or through managed cloud sessions. Kiro supports multiple foundational models, including Anthropic Claude (Opus, Sonnet, Haiku), DeepSeek, MiniMax, and an intelligent Auto mode that selects the model dynamically based on task complexity, latency, and cost. Kiro differentiates itself through its emphasis on automated reasoning, property-based testing (PBT), and open integration standards. Instead of solely relying on standard unit tests, Kiro tests requirements for contradictions and runs property tests to validate edge cases across all input spaces. It conforms to open ecosystem standards including the Agent Client Protocol (ACP), Model Context Protocol (MCP), AGENTS.md, and Skills.md, and offers headless CLI execution for CI/CD automation. Commercial access is managed through a credit-based pricing model with multipliers scaled to the chosen model tier, supported by AWS enterprise-grade security features such as IAM/SSO authentication, IP indemnity, and centralized governance controls.
Spec-driven development — Converts natural language prompts into formal requirements, architectural designs, and sequenced task checklists before code generation. - Property-based testing — Automatically generates property tests and checks requirements for logical contradictions to catch edge cases that bypass standard unit tests. - Multi-model selection and Auto-routing — Supports Claude Opus, Sonnet, Haiku, DeepSeek, and MiniMax, alongside an Auto mode that routes tasks based on complexity and cost. - Parallel subagent orchestration — Dispatches parallel subagents to execute separate components of a spec across local or cloud environments. - Headless CI/CD operations — Runs non-interactively in automated pipelines to perform code reviews, fix identified bugs, and open pull requests. - Standardized protocol support — Operates natively with Agent Client Protocol (ACP), Model Context Protocol (MCP), AGENTS.md, and Skills.md. - Enterprise AWS governance — Provides IAM and SSO authentication, centralized usage dashboards, cost controls, and IP indemnity.
Use cases
Use Case: Spec-Driven Feature Implementation When building a new microservice or module, a developer inputs core project requirements into Kiro. The tool generates an architectural design, creates task breakdowns, and orchestrates parallel subagents to scaffold directory structures, implementation code, and configuration files according to the spec.
Use Case: Automated CI/CD Code Review and Remediation Teams integrate the headless Kiro CLI into GitHub Actions or GitLab CI pipelines. Upon PR creation, Kiro analyzes the diff against architectural requirements, runs property-based tests against edge cases, and automatically commits fixes or appends structured review comments.
Use Case: Cloud Infrastructure and Configuration Authoring Cloud engineers use Kiro to draft and validate complex Terraform modules, container specifications, and AWS service integrations. Kiro parses existing infrastructure context and validates configuration properties prior to deployment.
Who it is for
Kiro is built for software engineers, cloud architects, and enterprise engineering teams seeking structured, spec-driven code generation with formal verification and AWS enterprise controls. It is less suited for developers looking for a lightweight, single-file completion snippet tool with zero project orchestration overhead.
Install guide
Installation Install the Kiro CLI using the official installation script: bash curl -fsSL https://cli.kiro.dev/install | bash
First Run Authenticate using your AWS Builder ID or supported enterprise SSO provider to initialize credentials and obtain platform credits: bash kiro auth login
Example Session Initialize a new spec-driven project session and run parallel agent tasks from the command line: bash # Generate a spec and start an interactive session kiro init --spec "Build a rate-limiting middleware in TypeScript with property tests" # Or run a headless cloud session for automated execution kiro run --cloud --autopilot
Amp — Amp (developed by AmpCode) is an autonomous coding agent platform and remote development harness designed to execute software engineering tasks…
Factory — Factory is an agent-native software development platform built around its autonomous coding agent, Droid. Developed by Factory AI, the system…
Devin — Devin is an autonomous AI software engineer developed by Cognition. It is designed to handle end-to-end software engineering workflows, including…
OpenHands — OpenHands (formerly OpenDevin) is an open-source autonomous AI software development agent and orchestration platform. It provides an extensible runtime environment…
Continue — Continue is an open-source AI code assistant and agent framework originally developed by the Continue team (continuedev) and recently acquired…
OpenCode — OpenCode is an open-source AI coding agent developed by Anomaly that operates across terminal environments, desktop applications, and IDE extensions.…