Kimi Code CLI: Terminal AI Coding Agent for Next-Gen Agents
Kimi Code CLI is an AI coding agent that runs in your terminal. It reads and edits code, runs shell commands, searches files, fetches web pages, and autonomously decides next steps based on feedback. It ships as a single binary with millisecond startup, features a purpose-built TUI, supports video input, AI-native MCP configuration, a rich plugin ecosystem, subagents for parallel tasks, lifecycle hooks, and ACP integration with editors like Zed and JetBrains. Ideal for developers seeking a powerful, extensible AI coding companion.
Kimi Code CLI is an AI coding agent that runs directly in your terminal. It can read and edit code, execute shell commands, search files, fetch web pages, and autonomously decide the next step based on feedback. It works out-of-the-box with Moonshot AI's Kimi models and supports other compatible providers.
Kimi Code CLI 是一个直接在终端中运行的 AI 编码代理。它可以读取和编辑代码、执行 shell 命令、搜索文件、获取网页内容，并根据反馈自主决定下一步操作。开箱即用 Moonshot AI 的 Kimi 模型，也支持配置其他兼容的模型提供商。
Traditional terminal AI tools often require Node.js setup, suffer from slow startup times, and have clunky interfaces. Kimi Code CLI addresses these with a single-binary distribution (no runtime dependencies), millisecond TUI startup, and a purpose-built terminal UI optimized for long agent sessions.
传统终端 AI 工具往往需要安装 Node.js、启动缓慢，且界面不够友好。Kimi Code CLI 通过单二进制分发（无运行时依赖）、毫秒级 TUI 启动、以及专为长时间代理会话优化的终端界面，解决了这些问题。
It supports video input — drop a screen recording for the agent to watch and act on. Built-in subagents (coder, explore, plan) run in isolated contexts for parallel work. Lifecycle hooks gate risky tool calls. The plugin ecosystem lets you install skills, MCP servers, and data sources from a marketplace or GitHub. AI-native MCP configuration avoids manual JSON editing.
支持视频输入——可以放入屏幕录制让代理观看并采取行动。内置子代理（编码、探索、规划）在独立上下文中并行工作。生命周期钩子可以拦截风险工具调用。插件生态允许从市场或 GitHub 安装技能、MCP 服务器和数据源。AI 原生的 MCP 配置免去了手动编辑 JSON。
Kimi Code CLI is a single-binary application built on top of pi-tui for the TUI layer. It speaks the Agent Client Protocol (ACP), enabling editors like Zed and JetBrains to drive sessions over stdio via the kimi acp subcommand. The agent uses Moonshot AI's Kimi models (or compatible APIs) for reasoning, and manages subagents, MCP servers, and hooks through a unified configuration.
Kimi Code CLI 是一个单二进制应用，TUI 层基于 pi-tui。它使用 Agent Client Protocol (ACP) 协议，使 Zed、JetBrains 等编辑器能通过 kimi acp 子命令在 stdio 上驱动会话。代理使用 Moonshot AI 的 Kimi 模型（或兼容 API）进行推理，并通过统一配置管理子代理、MCP 服务器和钩子。
Install on macOS/Linux: curl -fsSL https://code.kimi.com/kimi-code/install.sh | bash on Windows (PowerShell): irm https://code.kimi.com/kimi-code/install.ps1 | iex. After install, run kimi in a project directory, then /login to authenticate. Try a task like: 'Take a look at this project and explain its main directories.'
macOS/Linux 安装：curl -fsSL https://code.kimi.com/kimi-code/install.sh | bash Windows (PowerShell)：irm https://code.kimi.com/kimi-code/install.ps1 | iex。安装后，在项目目录运行 kimi，然后 /login 登录。尝试任务，例如：“查看这个项目并解释其主要目录。”
To use Kimi Code CLI within Zed, add the following to ~/.config/zed/settings.json:
{
"agent_servers": {
"Kimi Code CLI": {
"type": "custom",
"command": "kimi",
"args": ["acp"],
"env": {}
}
}
}
Then open a new conversation in Zed's Agent panel. Similar setup works for JetBrains via the ACP plugin.
要在 Zed 中使用 Kimi Code CLI，将以下配置添加到 ~/.config/zed/settings.json：
{
"agent_servers": {
"Kimi Code CLI": {
"type": "custom",
"command": "kimi",
"args": ["acp"],
"env": {}
}
}
}
之后在 Zed 的 Agent 面板中打开新对话即可。JetBrains 通过 ACP 插件有类似配置。
Ideal for developers who want an AI assistant deeply integrated with their terminal workflow — code review, refactoring, issue debugging, or generating documentation. Best on macOS/Linux; Windows requires Git Bash. Requires network access to Moonshot AI APIs or compatible providers. Without internet, core AI features are unavailable. Video input and subagents are available in supported models.
适合希望 AI 助手深度集成到终端工作流的开发者——代码审查、重构、问题调试或生成文档。在 macOS/Linux 上体验最佳；Windows 需要 Git Bash。需要网络连接以访问 Moonshot AI API 或兼容提供商。无网络时，核心 AI 功能不可用。视频输入和子代理仅在支持的模型中可用。