GV

GeezerVison

The World Wide Web, restored to 800 pixels of glory.

Home|About|Gaming cabinet|Archives|How it works|Starter sites

On this site

  • Download Kiro Crew
  • Downloads
  • What’s New
  • Match the model to the work with Auto
  • Kiro bots: delegate recurring work
  • Kiro for Students: 41 more universities eligible
  • About Kiro
  • Agents
  • IDE
  • CLI
  • Web
  • Mobile
  • Crew
  • Pricing

This view

  • Original page
  • New address

GeezerVison is showing kiro.dev — Kiro: Ship quality software

Tap to view full image

Try Kiro Crew in action

LocalRemote
⌘K — Search for anything…
2
Sessions
New
Older Sessions
Monthly event quota per site
Add a monthly event quota per site, and drop events once a site goes over it. Split it into research, build and review. Don't merge until I approve.
Spawned 1 sub-agentresearch · plan the quota check
Readingest path · usage counters · site settings
Createdruns/event-quota/research.md
The research plan is ready in runs/event-quota/research.md.
  • Where to check: at ingest, before an event is stored, so a site over its quota costs nothing to process.
  • What to count: stored events per site per calendar month in UTC, from the counter the usage page already reads.
  • Where to set it: a new quota field in site settings. Empty means no limit, so existing sites are unchanged.
One question before the build starts: should events sent through the import API count toward the quota, or only live tracking?
auto · 2.07 credits · 1m 42s
Live tracking only. Imports are backfills and shouldn't count. Go ahead with the build and review.
Spawned 1 sub-agentbuild · on feat/event-quota
Editedingest · drop and count events over the quota
Editedsite settings · monthly quota field
Rantests · unit 5 passed · integration 2 passed
Spawned 1 sub-agentreview · diff against main
Createdruns/event-quota/review.md
Ready to merge after your approval. The reviewer reran both suites: unit 5 of 5, integration 2 of 2.
  • Quota: set per site and counted per calendar month in UTC. Imports don't count, as you asked.
  • Over the quota: new events are dropped and counted, so the site owner can see what was lost.
  • Top finding (medium): domains are compared case-sensitively, so a quota saved for a domain with capitals never applies.
Waiting for your approval before merging into main.
auto · 2.07 credits · 1m 42s
defaultanalytics
auto

Crew

Persistent agents that pick up work and keep going while you’re away.

Download Crew

Orchestrate agent swarms that run for hours

Kiro Crew watches your backlog, codebase, or CI and acts when work appears.

  • Breaks work into smaller tasks
  • Proactive bots do specialized work
  • Brings the results back for review
  • Records what it did
Meet Kiro Crew

Hand off work and come back to progress

Cloud sessions move long-running tasks to the cloud with parallel agents that can run for hours, so work continues when you close your laptop.

  • Deterministic workflows
  • Isolated steps, no model shortcutting
  • Quality enforced at every step
Explore Kiro Web

Works how you do

Run Kiro wherever work starts, so a failing build turns into a fix.

  • Headless in CI
  • From Slack
  • On a schedule
  • On remote machines
Run Kiro in CI/CD

Quality enforced at every step

Workflows split long tasks into checked steps, so agents can’t skip a test that failed.

How workflows work

Don’t repeat yourself

Agents get the context behind the code, including past decisions, project standards, and previous work.

  • Persistent memory across sessions
  • Shared team knowledge
  • Corrections saved as reusable guidance
How memory works

Verified against your intent

Specs capture what you meant before code is written and verify results after.

  • Property-based tests across hundreds of cases
  • Requirements analysis that catches gaps before design
  • Bugfix Specs that fix the root cause without regressions
How specs work

Building blocks for automated software delivery

Kiro gives you the primitives to turn the software lifecycle into a loop you direct as autonomously as you’re comfortable with: persistent agents pick up signals and tasks, build and verify in sandboxes, flag what needs review, and learn from each session.

Frontier engineeringFrontier teams
Cloud sessionLocal session in CrewOptional step
Kiro CrewOne sandboxed run

01Signals

  • SlackThreads, DMs, and mentions
  • Microsoft TeamsChannels and chats
  • GitHubIssues and PR comments
  • GitLabIssues and MR comments
  • Meetings via MCPCall notes and transcripts
  • 100s of other appsPowers and MCP servers

02Triggers

  • Channel watchHeartbeat, hourlyLocal session in Crew
  • Scheduled scanAutomation, every 6 hoursCloud session
  • Explicit requestsMention, DM, or webhookLocal session in Crew
  • CI triageHeadless CLI on failureCloud session

03Triage

  • TriageDedupe, size, and route
    • Small: build it now
    • Big: spec it for a human

04Build

From Triage: Small

  • SandboxFresh clone, isolated
    • Reproduce
    • Smallest fix
    • Tests and lint
    Local or cloud
  • EvidenceProof the fix works
    Before
    After
    • 48 tests passed
    • Lint clean

    Page load 412 ms → 188 ms

05Review

From Build: Open PR

  • Pull requestGitHub or GitLab
    Push fixes
  • VerificationProperty tests and steering
  • WorkflowsAdversarial review agents
  • Human reviewOwner or requester

06Production

From Review: Merge

  • Merge and deployYour pipeline, your gates
  • Observability signalsErrors, metrics, and alerts via MCP
Track the big ones

From Triage: Track the big ones

Issue trackerSpecs and follow-ups for a human
Production signals start the next run

01Signals

  • SlackThreads, DMs, and mentions
  • Microsoft TeamsChannels and chats
  • GitHubIssues and PR comments
  • GitLabIssues and MR comments
  • Meetings via MCPCall notes and transcripts
  • 100s of other appsPowers and MCP servers

02Triggers

  • Channel watchHeartbeat, hourlyLocal session in Crew
  • Scheduled scanAutomation, every 6 hoursCloud session
  • Explicit requestsMention, DM, or webhookLocal session in Crew
  • CI triageHeadless CLI on failureCloud session
Kiro CrewOne sandboxed run

03Triage

  • TriageDedupe, size, and route
    • Small: build it now
    • Big: spec it for a human
Small

04Build

From Triage: Small

  • SandboxFresh clone, isolated
    • Reproduce
    • Smallest fix
    • Tests and lint
    Local or cloud
  • EvidenceProof the fix works
    Before
    After
    • 48 tests passed
    • Lint clean

    Page load 412 ms → 188 ms

Track the big ones

From Triage: Track the big ones

Issue trackerSpecs and follow-ups for a human
Open PR

05Review

From Build: Open PR

  • Pull requestGitHub or GitLab
    Push fixes
  • VerificationProperty tests and steering
  • WorkflowsAdversarial review agents
  • Human reviewOwner or requester
Merge

06Production

From Review: Merge

  • Merge and deployYour pipeline, your gates
  • Observability signalsErrors, metrics, and alerts via MCP

↺Production signals start the next run

Built for how frontier teams ship

Frontier teamsCase studies

The frontier models that matter

Use Claude, GPT, and open-weight models like DeepSeek and Qwen, or let Auto route each task to the right model.

Learn about models

Built on open standards

Bring the Kiro agent to your tools with ACP, and use AGENTS.md, Skills, MCP, and Open VSX extensions.

Enterprise-ready

Built and run by AWS, with SSO, governance, compliance, IP indemnity (with model exclusions), and predictable pricing.

Kiro for enterprise

Trusted by engineers and enterprises worldwide

Read customer case studies
  • Mondelez
  • Hughes Network Systems
  • Rackspace
  • FINRA
  • SmugMug
  • Appian
  • Netsmart
  • Siemens
  • Iress
  • NTT Data
  • Mondelez
  • Hughes Network Systems
  • Rackspace
  • FINRA
  • SmugMug
  • Appian
  • Netsmart
  • Siemens
  • Iress
  • NTT Data
  • “As a two hundred person team, we're tackling more complex features now with greater confidence, which is crucial when you're handling tens of billions of memories for more than 100M customers.”

    Don MacAskill

    CEO & Founder, SmugMug

  • “As a startup co-founder and CTO, time is the most important resource. Kiro justifies the use of my time for developing our business critical assets in-house.”

    RO

    Rolf Koski

    CTO & Co-Founder

  • “Kiro has helped my teams achieve a 35% uplift in productivity and a 17% improvement in value versus failure demand, accelerating roadmap delivery. Kiro is rapidly evolving our SDLC and fundamentally changing what we deem viable, allowing us to be far more aspirational with the projects our teams drive.”

    Kieron Wallace

    Head of Engineering Wealth UK, Iress

  • “I’ve been blown away by Kiro’s capabilities. The agentic experience is really transformative. From the multimodal inputs that understand context to the complete lifecycle control within the IDE, it feels like I’m working with a senior developer.”

    VI

    Vivek Velso

    Cloud Engineering Lead

  • “Most tools are great at generating code, but Kiro gives structure to the chaos before you write a single line.”

    FA

    Farah Abdirahman

    Cloud & AI Engineer

  • “In a very short time, our team transitioned from resistance to excitement, with developers enthusiastically sharing their achievements made possible by Kiro.”

    Brian Lichtle

    CTO ITS, Rackspace

  • “In roughly two days, I built a secure file sharing application from scratch. By simply sharing my requirements with Kiro, I was able to create a fully secure application that incorporates encryption and various security coding practices—no additional prompts needed.”

    IH

    Ihor Sasovets

    Lead Security Engineer

  • “Kiro is a strong ally for startups. It naturally turns overlooked docs and specs into robust assets, making growth smoother and future scaling more effective.”

    KE

    Kento Ikeda

    Founder & Engineer

  • “Kiro has been phenomenal in helping us adopt an “AI-driven development lifecycle” mindset. The ability to prototype and build simplified proofs-of-concept in a matter of hours is a game-changer.”

    Sahil Gupta

    Director, Appian

  • “I use Kiro for everything - drafting new Terraform modules, tweaking container setups, even writing down random AI ideas at 2 am. But more than anything, it supports how I learn. I’m endlessly curious, and Kiro helps me stay in that learning loop—tinkering, breaking things, fixing them, and then allowing me to share what I learn back with the community.”

    AD

    Adit Modi

    Solution Architect

  • “Kiro's AI-powered spec workflow turned rough ideas into polished, production-ready specifications in minutes, and is helping redefine what's possible with a lean team. A production-ready AWS serverless guardrail service featuring authentication, real-time dashboards, and complete infrastructure—a project traditionally requiring 3-4 months with multiple developers—was completed in 2 weeks by a single architect.”

    Fabian Fischer

    IT Enterprise Architect, Siemens

  • “I often forget to add unit tests, or update documentation when pushing changes, but with Kiro I can create a hook and it will automatically run those tasks in the background for me, never having to think twice.”

    DA

    Darya Petrashka

    Senior Data Scientist

  • “As an open source developer, I typically don’t build games, but last night I was able to create one using Kiro. I didn’t have to think too hard about implementation since Kiro handled the logic and enabled me to write code quickly and accurately.”

    MI

    Michael Walmsley

    Serverless Application Architect

  • “Kiro lowers the barrier to entry, whether that be, you don’t even know how to code, you’re new to coding, or you’re like me and you’re just learning a new tech stack. If something doesn’t make sense to you, or you feel like you are asking a silly question, ask Kiro anyway!”

    DA

    Danielle Heberling

    Founding Infrastructure Engineer

  • “Kiro comes equipped with all the best practices that I would put in my specs, and builds me the application I want, faster.”

    BL

    Blake Romano

    Senior Software Engineer

  • “Thanks to Kiro’s spec-driven development, I was able to go from concept to working prototype in a single weekend.”

    DA

    Dakota Lewallen

    Software Engineer

  • “In just four lines into a spec, Kiro was able to write user stories like a product manager and capture so many details that I didn’t even need to mention, saving me several days of work.”

    SA

    Sathiesh Veera

    Principal Software Engineer

  • “Kiro isn’t just speeding up my work – it’s making me write better code, think more systematically about architecture, and deliver higher quality results while automated workflows happen seamlessly in the background.”

    AN

    Anuj Tyagi

    Developer

  • “Kiro is just more than code. Aside from just listening to prompts and doing things, it can actually work on tasks based on predefined triggers, making the development experience more efficient.”

    SR

    Sreekesh Iyer

    Software Engineer

Get started fast with quick guides

Run Kiro in the cloud

Learn what a cloud session is, when to use one, what keeps it safe, and how to keep your configuration consistent between local and cloud.

{Blog}

Match the model to the work

{Guide}

Your first project

Pick the app that’s right for you

Loading image...
Kiro Crew

Crew

Download Kiro Crew
Loading image...
Kiro Web

Web

Sign up / in
Loading image...
Kiro Mobile

Mobile

Request early access
Loading image...
Kiro IDE

IDE

Download IDE
Loading image...
Kiro CLI

CLI

curl -fsSL https://cli.kiro.dev/install | bash

Common questions

What is Kiro?

Kiro is an agentic AI engineering system available via an [IDE](/ide/), [CLI](/cli/), [web interface](/web/), [mobile app](/mobile/), and [Kiro Crew](/crew/) that helps developers and teams do their best work. From quick fixes to complex tasks, Kiro works alongside you to turn prompts into executable specs, validate code correctness to find bugs unit tests miss, and build across large codebases with parallel agents that learn from every session—so what you build is exactly what you want and ready to share with your team. Kiro agents help you solve challenging problems, automate tasks, and keep working in the cloud when you step away.

What programming languages does Kiro support?

Kiro supports a variety of programming languages that developers use in their day to day work. This list includes, but is not limited to Python, Java, JavaScript, TypeScript, C#, Go, Rust, PHP, Ruby, Kotlin, C, C++, shell scripting, SQL, Scala, JSON, YAML, and HCL.

What is an AI agent?

An agent is an AI system that can plan and execute multi-step tasks with some degree of autonomy. Unlike simple chatbots that respond to individual prompts, agents can break down complex tasks, make decisions about how to proceed, and take actions to accomplish goals.
Browse all FAQs

Build something real in minutes

Get started for free

Download Kiro CrewDownloads
Install CLI
curl -fsSL https://cli.kiro.dev/install | bash
You are visitor 01402503 — GeezerVison does not store the pages you look up.
Best viewed with Netscape Navigator 4.7 or Internet Explorer 6 at 800×600. No cookies required.