Loading image...Kiro

Product

  • About Kiro
  • Agents
  • IDE
  • CLI
  • Web
  • Mobile
  • Crew
  • Pricing
  • Downloads

For

  • Enterprise
  • Startups
  • Students

Community

  • Overview
  • Ambassadors
  • Case studies
  • Discord
  • Events
  • Powers
  • Shop
  • Showcase

Resources

  • Docs
  • Blog
  • Changelog
  • FAQs
  • Videos
  • Report a bug
  • Suggest an idea
  • Billing support

Social

Site TermsLicenseResponsible AI PolicyLegalPrivacy PolicyCookie Preferences
Kiro
  • Agents
  • Enterprise
  • Pricing
  • Docs
SIGN INDOWNLOADS
Kiro

Get Started

InstallationAuthenticationYour first project

Models

OverviewAvailable modelsReasoning effortAWS GovCloud (US) models

Features

How Kiro worksLocal command sandboxACP integrations
Specs
Steering
Hooks
MCP
Permissions
Custom agents
Workflows
Agent Skills
Powers
Cloud sessionsMemoryCompactionKiroignoreCheckpoints and rewind
Built-in tools
Configuration scopes

IDE 1.x

What's new in 1.0
Setup & First Run
Editor
Chat
Experimental
Troubleshooting0.x reference

CLI

Setup & First RunChoose an agent harness
Terminal UI
Chat
Fullscreen modeVoice modeHeadless modeACPAuto complete
Experimental
Troubleshooting
Upgrade from CLI 2.x
2.x reference

Crew

Quick startInstallationRunning 24/7
Chat
Agent Capabilities
Features
Interfaces
Apps
System & storageConfigurationSecurityTroubleshooting

Web

Setup & First RunIdentity Center
Connect your repositories
GitHub
GitLab
Working with the agent
Chatting with the agent
Creating tasks
File explorer
Autonomous modeScheduleConfiguration Sync
Sandbox

Mobile - Preview

Overview

Commands and Reference

CLI commandsSlash commandsBuilt-in toolsExit codesSettings

Billing

OverviewManaging your subscriptionUpgrading your planDowngrading your planCancelling your planPurchasing add-on creditsManaging your paymentsManaging usage notificationsManaging your taxesContacting billing supportDeleting your accountRelated questions

Enterprise

ConceptsOnboarding quickstart
Connecting your identity provider
Deployment optionsSubscribe your teamManage subscriptions
Governance
Monitor and track
SettingsManaged updatesTLS inspection compatibilityBillingIAMSupported regions
AWS GovCloud (US)

Privacy and Security

OverviewData protectionCode referencesCompliance validationInfrastructure securityIAM permissionsFirewalls, proxies, and data perimetersVPC endpoints (AWS PrivateLink)

Guides

OverviewRunning Kiro in the cloud
Language support
Learn by playing

Migration

Migrating from Q DeveloperMigrating from VSCodeUpgrading from Q CLI
  1. Docs
  2. Web
  3. Working with the agent
  4. Chatting with the agent
View as Markdown

Chatting with the agent

View as Markdown

You can chat with the Kiro Web agent to discuss approaches, ask questions, and provide guidance before and during task execution.

Before creating a task

Use chat to:

  • Discuss different implementation approaches
  • Clarify requirements or constraints
  • Get the agent's input on technical decisions

The agent has access to web search, learnings from previous code reviews, and context from other tasks to provide informed responses.

During task execution

Once a task is created, you can continue chatting to:

  • Steer the agent on the implementation approach
  • Provide additional requirements or constraints
  • Ask the agent to do more work after reviewing initial results

Any comments or steering will update the scope of the current task. You cannot create a second task in the same session — to work on a different task, start a new session.

Parent chat and workflow steps

A background workflow keeps the main conversation available while its nested step sessions run. Use the two conversation surfaces deliberately:

  • Parent chat: ask the main agent to interpret progress, solve a broader issue, change the overall direction, or summarize completed evidence.
  • Selected step: open Workflows, select a node, and use the state-specific field: Answer this step when paused, Steer the agent while running, or Message this step's agent otherwise.

A direct step message does not necessarily interrupt an active tool call. It applies according to the step's current state and turn boundary. When a step explicitly pauses for input, reply in that step session rather than only in the parent chat.

Page updated: September 30, 2026
Working with the agent
Creating tasks