Changelog

This release lets you talk to Kiro instead of typing: /voice dictates your prompts with on-device transcription. Spec runs gain a review screen for reading phase documents and staging line comments, and AGENTS.md files now load as steering context from anywhere in your workspace tree. Enterprise administrators should note that cloud sessions are now opt-in, so organizations that never set the toggle will find them disabled.

Speak your prompts instead of typing them. Start a recording with /voice or by holding Space, and partial text streams into the prompt as you talk until you press Enter or it auto-stops on silence. By default your speech is transcribed on your own machine with Whisper, so no audio leaves your computer and no cloud API key is required. The model downloads once, behind an explicit confirmation, on first use. Learn more ->

Cloud sessions are now disabled unless an administrator explicitly turns on Cloud Sessions in the Kiro console. Because an unset toggle now resolves to disabled, organizations that never configured it will find cloud sessions unavailable until an admin opts in. Learn more ->

Press Ctrl+X at a spec phase checkpoint to read the phase document in place and stage line comments on the parts you want changed. When you answer the checkpoint question, every staged comment goes to the agent as one revision request, so reading a document and asking for changes stays in a single screen. Learn more ->

AGENTS.md files now load as steering context from anywhere in your workspace tree, not just the workspace root and ~/.kiro/steering/. Put an AGENTS.md next to the code it describes, and the agent picks it up alongside your other steering files. Learn more ->

This release introduces cloud sessions in preview: run your session in a managed cloud sandbox with --cloud, disconnect while the agent keeps working, and resume from any machine. The slash command menu also gains substring matching, and typing in the TUI stays smoother during rapid input.

Cloud sessions let you run Kiro CLI in a managed cloud sandbox instead of on your machine. Start one with kiro-cli --cloud (or kiro-cli chat --cloud), attach repositories with --repo or the /repo picker, then disconnect while the agent keeps working and resume later from any machine with --resume-id. Cloud and local sessions appear together in the session picker. Learn more ->

For organizations using IAM Identity Center, administrators manage Cloud Sessions with the Cloud Sessions (Preview) setting, formerly Kiro Web (Preview). Organizations that enabled the prior setting remain enabled; new organizations must opt in before users can start Kiro Web sessions.

This version brings Cloud Sessions to Agent Focus Mode as a preview and gives the sessions rail a collapsible icon-only view, attention cards you can answer in place, and back/forward navigation. It also pins the agent's Windows terminal to PowerShell, restores Always Allow for commands it was quietly forgetting, and fixes the macOS updater downloading every update twice.

Cloud Sessions are now available in Agent Focus Mode as a preview. Start a session that runs in the cloud alongside your local ones, pick the repositories it works against, and keep those repositories visible below the chat. Cloud sessions use the same agent and model selectors as local sessions. Learn more ->

Collapse the sessions rail to an icon strip to give the chat more room, with every session still one click away. When a session needs your input, an attention card appears next to the chat with the question and the agent's options as buttons, so you can answer it without switching over. The attention panel never opens on its own, and won't raise itself for the session you already have on screen. And you can move back and forward across sessions and settings from the titlebar (Cmd/Ctrl+[ and Cmd/Ctrl+]). Learn more ->

You can now bring your personal .kiro configuration into Kiro Web. From Settings > Configuration Sync, upload local Steering, custom agents, Hooks, Skills, Powers, or MCP settings for use in every cloud session. Learn more ->

Select or drag one supported top-level folder, such as skills/ or agents/, from your local .kiro directory. Kiro checks each file against its cloud copy before uploading. New files are selected by default, while updates that replace cloud content require your input. Steering files and MCP servers include a side-by-side comparison.

Uploaded items appear on their matching settings pages, where you can view, edit, download, or delete each one. Steering, Skills, Agents, Hooks, and Powers each have their own page under Settings, and uploaded MCP servers appear with your MCP server settings. Catalog Powers are installed by name, and custom Powers upload as text files.

The first feature release since Kiro Crew went open source brings a real browser the agent can operate, eight new built-in apps, a native Windows desktop build, Korean and Japanese interfaces, and setup that no longer assumes Slack, alongside several hundred fixes from the first weeks in the open.

Flip one switch in Settings and the agent can operate a real browser: navigate, click, type, and fill forms, with the live view streaming into the dashboard's Browser panel. Installation happens for you and recovers on its own, and the agent can also serve browser work from the native embedded view. Learn more ->

The App Store adds Spec Builder (spec-driven development), Ops Mission Control (an autonomous ops first responder), Auto-Improvement (measurement-first self-improvement), Meetings (live transcription with structured notes and action items), Papyrus (a LaTeX editor with an AI co-author), Mochi and Crew Companion (desktop companions), and PPTX Maker (describe a deck in chat, get a real .pptx back). Every app is opt-in: install it from the App Store and enable it before it does anything. Learn more ->

The desktop app now ships as a native Windows installer with an integrated titlebar, launcher fixes, and a configurable sandbox tier for agent subprocesses, with skills, usage tracking, and build tooling following Windows conventions. Learn more ->

kirocrew setup no longer asks for Slack tokens. The wizard finishes on the dashboard and points at the full set of chat channels; walk through Slack credentials only when you ask for them with kirocrew setup --slack. Learn more ->

Korean and Japanese join the dashboard, bringing the interface to twelve languages. Bold, italic, and strikethrough now render correctly in CJK prose. Learn more ->

This version supports Powers packaged in the open Agent Plugin format and reworks the Agent Focus Mode session rail with pinning, Kiro CLI hand-off, and in-app updates. It also moves Kiro to Code OSS v1.109.5, improves how spec hovers and MCP installs handle untrusted workspace content, and adds bulk session migration and steadier error recovery.

Install powers aligned with the open Agent Plugin format from a local folder or GitHub URL. Kiro now supports plugins that bundle skills and MCP making powers easier to share across compatible agent tools.

Pin a session with Pin Session to keep it at the top of its section. Resume an existing session in Kiro CLI with Open with Kiro CLI. When a new version of Kiro is available, an update row appears in the rail footer, and the settings pane gains a Check for Updates button. Spec documents and task rows in the context panel now open directly in the editor. Learn more ->

Kiro Crew is now open source. It is a personal AI agent that runs on your own machine, driving kiro-cli over the Agent Client Protocol. Install it, sign in once, and it is yours: no server to rent, no account to create, and your conversations, memory, and files stay on your disk.

Chat with the same agent from a web dashboard, a native desktop app, a terminal CLI (kirocrew chat), or bots for Slack, Discord, Telegram, Microsoft Teams, Webex, WeCom, and WeChat. Every channel shares the same memory and tools, so you can start something at your desk and follow up from your phone. A dashboard session can even hand off to a Slack thread and stay in sync both ways. Learn more ->

Hand the agent a spec with kirocrew run TASK.md and it decomposes, executes, tests, and retries, designed for 10+ hour runs with disk checkpoints so a crash resumes where it stopped. Schedule recurring jobs with cron, split work across parallel sub-agents with kirocrew spawn run, or let a webhook wake the agent when CI or an alert fires. Learn more ->

Preferences, project context, and conversation history persist across sessions and are searched by keyword and by meaning, with embeddings running locally so nothing leaves your machine. Correct the agent once and the lesson sticks in every future session. A Knowledge Library lets you ingest your own documents and code into a searchable personal knowledge graph. Learn more ->

Extend the agent through an in-dashboard App Store with TypeScript and Python SDKs, six apps in the box (including Auto Research, Code Review Sage, and Issue Radar), and Skills: plain markdown files that teach the agent a workflow with no code and no rebuild. Any MCP server can be discovered, probed, and enabled from the dashboard. Learn more ->

The agent can read a native application through the accessibility layer and operate it: press buttons, fill fields, scroll, and drag, reaching work that has no web UI. Actions are delivered to the target app so your mouse pointer never moves by accident, password fields are never read, and every call is written to the audit log. Off by default and macOS-only in this release; enable it in Settings > Computer Use. Learn more ->

An opt-in operating-system sandbox confines kiro-cli subprocesses on Linux and macOS. Layered controls include 137 built-in denied-command patterns that hold even when auto-approval is on, credential redaction, and a two-level governance ceiling enforced at Kiro Crew's own tool gate, stored where the agent can neither read nor write it, so the agent cannot widen its own limits. Learn more ->

OpenAI reduced prices for GPT‑5.6 Terra and Luna, and we’re passing those savings on to Kiro customers. Updated credit multipliers reflect OpenAI’s new pricing and are live now.

• GPT‑5.6 Luna drops from 0.6x to 0.1x
• GPT‑5.6 Terra drops from 1.2x to 1.0x
• GPT‑5.6 Sol remains unchanged at 2.4x

GPT‑5.6 Sol, Terra, and Luna are available with experimental support to Kiro Pro, Pro+, Pro Max, and Power customers. See the original launch post for full details on GPT‑5.6 in Kiro.

This release introduces tangents: branch into side-conversations that inherit your full conversation history, explore freely, then jump back exactly where you left off. /context also gains a per-tool token breakdown so you can see what's consuming your context window.

Need to look something up mid-task without polluting your main thread? /tangent branches into a side-conversation that inherits everything discussed so far — explore freely, then jump back to exactly where you left off. Create named tangents with /tangent <name>, browse your whole conversation tree in a visual picker with /tangent ls, and jump straight back to the main conversation from any depth with /tangent root. Available in V3 mode (kiro-cli --v3). Learn more ->

This version adds a Kiro submenu to the editor context menu and an Ask Kiro to Fix quick fix on errors and warnings, plus a guided form for creating hooks. Kiro moves to Code OSS v1.108.2, and now replies in your IDE display language. Chat history also survives workspace folder changes, and several sign-in, trust, and approval issues are fixed.