Kiro
Kiro for specification-driven development: requirements, agents, credits and data rules for individual and organizational accounts.
Provider sources checkedWhat does Kiro do?
Who is this tool for?
- Developers and product teams with more complex requirements.
- AWS-oriented teams establishing a governed agent workflow.
Strengths
- Requirements and implementation can be assessed against the same criteria.
- Hooks and guidelines support recurring project tasks.
- Individual and organizational access have documented credit models.
Limitations and requirements
- Planning, refinement and hooks also consume credits.
- Supervised Mode is a review workflow, not a comprehensive access restriction.
- Automatically written requirements need business-level review.
Who is behind the product?
- Company
- Amazon Web Services
- Headquarters
- Not clearly established by the reviewed primary sources.
- Company founded
- Not clearly established by the reviewed primary sources.
- Product launch
- 14 July 2025: official IDE preview 0.1.
The first official changelog entry documents specification-driven IDE work. Kiro now describes a broader product family. AWS also names Kiro as a migration route for Q Developer capabilities; this does not mean every function or term is identical.
Suggested ways to use it
Suggested workflows or verification steps. These are not hands-on tests; for unverified products they are not a recommendation to use the service.
Suggested: role-based feature
- Specify roles, data and permitted actions as requirements.
- Review design and tasks for missing failure cases.
- Test implementation against the same criteria and inspect the diff.
Goal: A traceable connection between requirements and changes.
Suggested: recurring documentation check
- Define a narrow trigger and affected documents.
- Observe the hook in a test project first.
- Limit unnecessary runs and credit usage before ongoing adoption.
Goal: A controlled routine with measurable effort.
Pricing, allowances and data
What should you check before choosing a plan?
Free includes 50 monthly credits. Paid individual plans start with Pro at USD 20/month; further tiers are USD 40, 100 and 200. Additional credits cost USD 0.04 according to the current page. Team pools start with at least two seats at USD 20 each and 500 included credits per seat. Account for taxes and model-dependent consumption.
What should you check about data and permissions?
Free and paid individual content may be used for service improvement subject to opt-out; Enterprise content is not, according to documentation. Individual content is stored in US East, while inference may use other regions. Supervised Mode reviews changes but does not automatically limit reading or network access. Command permissions and sensitive environment variables need separate controls.
External reviews and available evidence
No GatwAI hands-on test. The assessment uses checked primary sources; it is not based on an independent original test of the current version.
Reported positives
- Specifications provide concrete criteria for your own acceptance review.
Reported cautions
- Workflows are suggestions, not test results. Provider claims do not establish a general success rate.
Product development and notable changes
IDE preview released
Version 0.1 introduced the documented requirements, design and task workflow.
Source for this change ↗
Which alternative fits which task?
Google Antigravity →
For another integrated agent and IDE workflow.
JetBrains Junie →
For agent work with JetBrains tooling and model selection.
Frequently asked questions
Does only code generation consume credits?
No. Specification work, refinement and hooks can also consume credits.
Is Supervised Mode a sandbox?
Do not equate them. It governs review; access, networking and trusted commands have separate rules.
