Describe a coding task in plain language and have the assistant carry it out step by step.
Connect external data sources through Model Context Protocol servers.
Dispatch focused coder, explore, or plan subagents on isolated pieces of a task.
Use lifecycle hooks to review or block actions before they run.
| moonshotai/kimi-code | 40rbidd3n/hydro0x01 | fastify/demo | |
|---|---|---|---|
| Stars | 234 | 229 | 227 |
| Language | TypeScript | TypeScript | TypeScript |
| Last pushed | — | — | 2026-06-30 |
| Maintenance | — | — | Active |
| Setup difficulty | easy | hard | moderate |
| Complexity | 2/5 | 4/5 | 3/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Installs as a single self-contained binary with no Node.js requirement to run.
Kimi Code CLI is a terminal-based AI coding assistant made by Moonshot AI. You run it from the command line inside a code project, and it can read and modify files, run shell commands, search through your codebase, and fetch web pages. It decides what to do next based on what it observes, so you describe a task in plain language and it tries to carry it out step by step. It is designed to work with Moonshot AI's Kimi models by default, but can also connect to other compatible AI providers. Installation is a single command that downloads a self-contained binary. There is no Node.js installation required to run it, and no global package setup. It works on macOS, Linux, and Windows. Once installed, you run "kimi" in your project folder, sign in with a Moonshot account or an API key, and start giving it tasks in natural language. Several features are highlighted in the README. The tool accepts video input, so you can drop in a screen recording to show the agent something that is difficult to type out. It supports Model Context Protocol servers, which are a way of connecting AI tools to external data sources, and these can be configured conversationally through a /mcp-config command rather than by editing configuration files by hand. Built-in subagents called "coder", "explore", and "plan" can be dispatched to work on focused pieces of a task in isolation. Lifecycle hooks let you run local scripts at defined points, for example to review or block a particular action before it happens. The terminal interface is described as fast to start and built for long working sessions. The project is MIT-licensed and open source. Contributing requires Node.js 24.15 or newer and the pnpm package manager.
A terminal-based AI coding assistant from Moonshot AI that can read, edit, and run commands in your codebase from natural-language instructions.
Mainly TypeScript. The stack also includes TypeScript, Node.js, pnpm.
Use, copy, and modify freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.