Sylwester Liljegren
f96771d302
High: visible tool content (e.g. a todowrite checklist) could be
highlighted in the DOM while the counter still reported "No results"
and navigation was disabled. toolText() previously only indexed
state.title for non-bash tools, independently of what
transcript-search-highlight.ts actually scans in the DOM. Rewrote it
to recursively collect every string leaf from state.input/state.metadata
plus state.output for any tool, so matching and highlighting draw from
one canonical, much more comprehensive notion of "the tool's text"
instead of two divergent ones. read/glob/grep/list remain excluded --
kilo-ui's context-tool-results.tsx confirmed they never render raw
input/output text, even expanded, so including it there would
reintroduce the same class of mismatch for those tools.
High: search only covered the initially-loaded message page (80
messages), silently missing older history in long sessions. MessageList
now auto-requests older pages (session.loadOlderMessages()) while a
search is active with a non-empty query, looping via reactivity on
hasOlderMessages()/loadingOlderMessages() until history is exhausted.
A new searchingHistory state surfaces this to the widget, which shows
"Searching earlier messages..." and withholds a final "No results"
until the whole session has actually been searched (a live match count
still updates progressively as pages load).
Medium: whole-word matching used plain \b, which only treats ASCII
letters/digits/underscore as word characters and silently breaks for
Cyrillic, Arabic, CJK, and similar text. Replaced with Unicode-aware
boundary lookarounds using \p{L}/\p{M}/\p{N} property escapes, with the
`u` flag applied to every compiled pattern.
New chat.search.searchingHistory i18n key added across all 20 locales.
(cherry picked from commit 47d4ebd73b)
English | 简体中文 | 繁體中文 | 한국어 | Deutsch | Español | Français | Italiano | Dansk | 日本語 | Polski | Русский | Bosanski | العربية | Norsk | Português (Brasil) | ไทย | Türkçe | Українська | বাংলা | Ελληνικά | Tiếng Việt
The open source coding agent for building with AI in VS Code, JetBrains, or the CLI.
Kilo Code is an AI coding agent that meets you everywhere you work: VS Code, JetBrains, and the CLI. It's open source with open pricing. You pick from 500+ models, switch between them mid-task, and pay the model provider's rate with zero markup. No API keys required to start.
Installation
Pick where you want to run Kilo.
VS Code
Install the Kilo Code extension directly, or grab it from the VS Code Marketplace. Create an account and you'll have access to 500+ models including GPT-5.5, Claude Opus 4.7, Claude Sonnet 4.6, and Gemini 3.1 Pro Preview, all at provider pricing.
CLI
# npm
npm install -g @kilocode/cli
# curl
curl -fsSL https://kilo.ai/cli/install | bash
# pnpm
pnpm add -g @kilocode/cli
# bun
bun add -g @kilocode/cli
# Homebrew (macOS / Linux)
brew install Kilo-Org/tap/kilo
# Arch Linux (AUR)
paru -S kilo-bin
Then run kilo in any project directory to start.
JetBrains
Install the Kilo Code plugin from the JetBrains Marketplace, or search "Kilo Code" in Settings → Plugins inside any JetBrains IDE.
Install the CLI from GitHub Releases (binaries)
Download the latest binary from the Releases page.
| Platform | Asset |
|---|---|
| Windows (most PCs) | kilo-windows-x64.zip |
| macOS (Apple Silicon) | kilo-darwin-arm64.zip |
| macOS (Intel) | kilo-darwin-x64.zip |
| Linux x64 | kilo-linux-x64.tar.gz |
| Linux ARM | kilo-linux-arm64.tar.gz |
Notes: x64-baseline is a compatibility build for older CPUs without AVX. musl is the statically linked build for Alpine or minimal Docker images without glibc. kilo-vscode-*.vsix is the VS Code extension package, not the CLI. Source code archives are for building from source.
Agents
Kilo ships with specialized agents you switch between depending on the task. You can also build your own custom agents.
- Code - The default. Implements and edits code from natural language.
- Plan - Designs architecture and writes implementation plans before any code gets written.
- Ask - Answers questions about your codebase without touching any files.
- Debug - Troubleshoots and traces issues.
- Review - Reviews your changes and surfaces issues across performance, security, style, and test coverage.
Learn more about agents and custom agents.
What it does
- Code generation from natural language, across multiple files.
- Inline autocomplete with ghost-text suggestions and tab to accept.
- Self-checking so the agent reviews and corrects its own work.
- Terminal and browser control to run commands and automate the web.
- MCP marketplace to find and wire up MCP servers that extend what the agent can do.
- 500+ models with mid-task switching, so you can match latency, cost, and reasoning to the job.
Autonomous Mode (CI/CD)
Run kilo run with --auto for fully autonomous operation with no prompts, built for CI/CD pipelines:
kilo run --auto "run tests and fix any failures"
--auto disables all permission prompts and lets the agent execute any action without confirmation. Only use it in trusted environments.
Documentation
For configuration and everything else, head over to the docs.
Contributing
Contributions are welcome from developers, writers, and everyone in between. Start with the Contributing Guide for environment setup, coding standards, and how to open a pull request. See RELEASING.md for the VS Code extension and CLI release process, and packages/kilo-jetbrains/RELEASING.md for the JetBrains plugin.
Please review our Code of Conduct before getting involved.
License
MIT. You're free to use, modify, and distribute this code, including commercially, as long as you keep the attribution and license notices. See License.
FAQ
Where did Kilo CLI come from?
Kilo CLI is a fork of OpenCode, enhanced to work within the Kilo agentic engineering platform.