Kiro Freemium

AWS's spec-driven agentic IDE that turns requirements into code and tests.

Kiro is an agentic IDE from AWS that writes three spec files — requirements.md, design.md and tasks.md — before it touches code. It uses a spec-driven workflow, hooks for automated checks, and runs on macOS, Windows and Linux with a free tier.

What is Kiro?

Most AI editors start typing the moment you ask. Kiro argues that is how you get plausible-looking code that solves the wrong problem. So it front-loads the thinking: you describe a feature, it produces a requirements doc in EARS notation, you edit it, it produces a design, you edit that, and only then does it generate a task list it works through.

It is built on Code OSS like VS Code, and AWS ships it as its own product rather than a plugin for an existing editor.

Product Features

  • Spec workflow: requirements, design and task documents generated as editable markdown files.
  • Agent hooks: trigger actions on file save — run tests, update docs, or lint a changed file automatically.
  • Steering files: repo-level rules that tell the agent your conventions, stack and forbidden patterns.
  • Autopilot and supervised modes: let it work through a task list unchecked, or approve each step.
  • MCP support: hook external tools and services into the agent context.
  • VS Code-compatible extensions via the Open VSX registry.

Product Highlights

  • Specs cut rework. Reviewing a 200-word requirements file is cheaper than reviewing 600 lines of generated code.
  • Hooks keep artifacts in sync. Auto-running tests on save catches agent mistakes before you notice them.
  • Steering files scale across a team. One shared conventions file applies to every agent run in the repo.

Where Kiro Falls Short

The spec process is heavy for small changes — writing three documents to fix a typo is annoying, and Kiro still nudges you toward it. Model choice is narrower than open competitors because it routes through AWS-hosted models. Free-tier agent requests are metered, so a long autopilot session can burn a day’s quota in an hour.

Common Use Cases

  • Greenfield features where requirements are fuzzy and you want them pinned down first.
  • Onboarding an agent to a legacy repo using steering files for house rules.
  • Automated test generation triggered on every save through hooks.
  • Building a REST service from an OpenAPI sketch with generated design docs.
  • Compliance-friendly workflows that need an auditable spec trail.

Quick Summary

  • Kiro is an AWS agentic IDE built around spec-driven development.
  • It generates requirements.md, design.md and tasks.md before coding.
  • It has a free tier with metered agent requests plus paid plans.
  • The spec overhead is a poor fit for tiny edits.