The walkthrough on this page covers 6 steps: 1. Install Kiro and open your first project 2. Write a feature spec from your prompt 3. Add correctness requirements 4. Run agents in parallel and steer them 5. Move to the terminal with Kiro CLI 6. Bring local setup into the cloud.

What platforms does Kiro support?

Kiro is available on Web App, MacOS, Windows and Linux.

What are the limitations of Kiro?

Spec driven workflow adds upfront writing and review that small fixes do not need. Correctness validation carries a learning curve and pays off mainly on well defined requirements. Cloud sessions do not inherit your local agents, MCP servers, hooks or steering until cloud config is set up.

Kiro

Name: Kiro
Author: Kiro

Agentic engineering that starts with executable specs, not loose prompts

Advanced

Visit Kiro Back to AI tools

Own Kiro? Verify ownership to get a dofollow link

FeaturedHand-picked by aiwizard

AI Angels400+ customizable AI girlfriends for NSFW chat, voice calls, and roleplay

Candy AIExperience the future of companionship with AI friends who chat, role-play, and connect with you on a human level.

ourdream.aiA vast AI roleplay platform with millions of characters for chat, images, video and voice

Screenshot of Kiro, Agentic engineering that starts with executable specs, not loose prompts

What is Kiro?

Kiro is an agentic engineering environment from AWS that converts prompts into executable specs, validates code against correctness properties, and runs parallel agents locally or in the cloud, with a matching CLI for terminal workflows.

Kiro is an AI development environment built around a simple premise: prompts alone are a weak foundation for serious software work. Instead of dropping a request into a chat box and hoping the output holds up, Kiro turns prompts into executable specs, structured documents that describe the feature, its correctness requirements and the tasks needed to ship it. The agent then works against those specs inside a codebase, and the resulting code can be validated against defined properties so that bugs which ordinary unit tests walk past become visible. The tool spans two surfaces. Kiro Crew is a workspace where multiple agent sessions can run in parallel, each working on its own task and feeding what it learns back into later sessions. Kiro CLI brings the same model to a terminal, with a V3 client that adds specs, expanded hooks and a reworked trust model, plus a lighter /lite mode for people who want the same features in a simpler shell. Sessions can run locally or in a cloud workspace, and cloud sessions keep running after you disconnect, so you can pick the thread back up later from the session list. In practice a typical run looks like this: describe the feature, let the agent draft the spec and a task list, approve or steer it while it works, then let correctness checks and hooks catch problems before review. Subagents, skills, MCP servers and a registry of editor extensions extend the base environment, and import guides exist for teams moving over from VS Code. Enterprise controls cover identity provider configuration and governance. Kiro is aimed at professional developers and engineering teams who already use AI coding assistants but want more structure and verification around the output. It sits closer to a spec driven development platform than to a pure autocomplete tool, so the value shows up most on larger codebases and multi step features rather than one line fixes.

How do you use Kiro?

1Install Kiro and open your first project
Follow the getting started guide to bring Kiro into an existing repository. The first project walkthrough covers the initial setup so the agent has a working context before you ask it to change anything.
2Write a feature spec from your prompt
Describe what you want to build. Kiro drafts a feature spec that turns the request into a structured, executable plan with tasks the agent can work through.
3Add correctness requirements
Attach correctness properties to the spec so the implementation gets validated against defined behaviour rather than only example based tests. This is where the workflow diverges most from plain chat driven coding.
4Run agents in parallel and steer them
Use the workspace to run several sessions at once, and type mid run to steer the agent without cancelling the task. Subagents let you split specialised work across the session.
5Move to the terminal with Kiro CLI
Install the CLI and start a chat session against your repository, choosing local or cloud execution. The command reference covers agents, specs, hooks and session switching.
6Bring local setup into the cloud
When you start a cloud session, move your agents, MCP servers, hooks and steering from your home directory configuration into the cloud workspace so remote runs behave like local ones.

Pros and cons

Pros

Specs give agent work a durable plan instead of a one shot promptAI
Correctness validation is designed to catch bugs that ordinary unit tests missAI
Parallel agent sessions in a shared workspace support larger, multi part featuresAI
Cloud sessions persist after disconnecting and can be resumed from a session listAI
CLI mirrors the editor experience, including a lighter mode for simpler terminalsAI

Cons

Spec driven workflow adds upfront writing and review that small fixes do not needAI
Correctness validation carries a learning curve and pays off mainly on well defined requirementsAI
Cloud sessions do not inherit your local agents, MCP servers, hooks or steering until cloud config is set upAI
Terminal usage assumes comfort with commands, flags and session managementAI

How much does Kiro cost?

Integrations

Kiro connects through MCP servers and supports editor extensions via a registry, with documented ACP support and a migration path for VS Code users.

Learn more

Features

Executable feature specs, correctness validation, parallel agent sessions in a shared workspace, subagents, skills, MCP support, an editor extension registry, a terminal CLI with cloud sessions and hooks, and enterprise identity and governance controls.

Learn more

Frequently asked questions about Kiro

How do you use Kiro?
The walkthrough on this page covers 6 steps: 1. Install Kiro and open your first project 2. Write a feature spec from your prompt 3. Add correctness requirements 4. Run agents in parallel and steer them 5. Move to the terminal with Kiro CLI 6. Bring local setup into the cloud.
What platforms does Kiro support?
Kiro is available on Web App, MacOS, Windows and Linux.
What does Kiro integrate with?
Kiro connects through MCP servers and supports editor extensions via a registry, with documented ACP support and a migration path for VS Code users.
What are the limitations of Kiro?
Spec driven workflow adds upfront writing and review that small fixes do not need. Correctness validation carries a learning curve and pays off mainly on well defined requirements. Cloud sessions do not inherit your local agents, MCP servers, hooks or steering until cloud config is set up.