QwenLM/qwen-code

An open-source AI coding agent that lives in your terminal.

What it solves

Qwen Code is an open-source AI coding agent designed to handle complex development tasks by automating coding, testing, and review processes. It eliminates vendor lock-in by providing an open framework that works with various AI models and integrates directly into your existing development environment.

How it works

It functions as an agentic harness that can be run in the terminal, as a desktop app, or via IDE plugins. You provide a goal, and the agent uses built-in skills, sub-agents, and persistent memory to execute tasks. It supports multi-protocol connections, allowing you to switch between different model providers (like OpenAI, Anthropic, or local models via Ollama) at runtime. It can also run as a background daemon to share sessions across different interfaces like web browsers and editors.

Who it’s for

It is for developers who want to automate coding workflows, perform batch code reviews, or manage complex multi-step programming tasks using an AI agent that can be controlled from the command line, IDE, or even chat platforms.

Highlights

  • Agentic core with built-in memory, sub-agents, and multi-agent coordination.
  • Multi-protocol support for various AI providers and local models.
  • Versatile interfaces including CLI, Desktop app, Web UI, and IDE plugins (VS Code, Zed, JetBrains).
  • Extensible via Model Context Protocol (MCP), custom scripts, and automated skill learning.
  • Built-in safety features like sandboxing and Git worktrees to isolate code changes.

Related

  • Project
  • Project
  • Project
  • Dispatch
  • Project