Out of nowhere, Amazon released its own AI IDE called Kiro, a fresh contender in the growing field of next‑generation developer tools.
Built on top of VS Code like its rivals Cursor, Windsor and Firebase Studio, Kiro sets itself apart with a spec‑driven workflow, Claude Sonnet 4.0 under the hood and a focus on disciplined software engineering. For now, it is free to use.
Another VS Code fork, but with a twist
Many AI editors rush you straight into auto‑generated code. Kiro deliberately slows the process by starting with clear specs before any code is written:
- requirements.md – user story and acceptance criteria
- design.md – architecture decisions, testing strategy, error handling
- implementation plan – merges both docs into discrete AI tasks
The flow feels slower yet far more intentional, ideal for teams that value clean and consistent code.
First impressions
Kiro’s UI is modern and uncluttered, but early adopters note a few rough edges:
- Occasional lag, likely due to launch‑week demand
- Missing small niceties such as chat checkpoints found in Cursor
- Closed‑source and limited to Claude models (for now)
Why Amazon built Kiro
Amazon is a major investor in Anthropic, having committed $8 billion so far. With Claude‑powered tools generating billions in revenue, an Amazon‑branded IDE makes strategic sense—especially after recent pricing friction at Cursor. Kiro offers more Claude for less money, challenging Cursor’s model head‑on.
Made for teams, not quick hacks
Everything about Kiro signals an enterprise focus:
- Planning before code generation
- Guardrails that reduce sloppy output
- Workflows that align with design reviews and team collaboration
For long‑term projects and multi‑developer teams, those guardrails are a benefit, not a burden.
Should you try Kiro?
If Cursor or Copilot already meets your needs, you do not have to switch. But if you run larger projects, manage multiple contributors or want tighter structure in your AI workflow, Kiro deserves a test drive—especially while it remains free.
Kiro is not racing on raw speed; it is competing on software‑engineering discipline.