# Qwen Code

An open-source AI coding agent that works from the terminal, an IDE, a desktop app, a browser or chat, with built-in memory, skills, subagents and MCP support.

Qwen Code is an open-source AI coding agent that works from the terminal, an IDE plugin, a desktop app, a web UI, or chat integrations such as Telegram, DingTalk, WeChat and Feishu. It ships with agentic features out of the box, including automatic memory, automatic skills, subagents, agent teams and MCP support, with no extra setup.

Both the framework and the underlying Qwen models are open source and developed together. It is multi-protocol, supporting OpenAI, Anthropic, Gemini and Qwen APIs, third-party providers, and local models through Ollama or vLLM, with models switchable at runtime.

It installs from a shell script on Linux and macOS, from PowerShell on Windows, or through npm (Node.js 22 or newer) or Homebrew, and is written in TypeScript under the Apache-2.0 license.

Website: https://qwenlm.github.io/qwen-code-docs/en/users/overview