Qwen Code

Active
GitHub TypeScript Apache-2.0

Description

An open-source AI coding agent that lives in your terminal, built by Qwen team with support for code generation, editing, debugging and multi-file operations.

Key Features

  • Multi-protocol multi-provider — supports OpenAI, Anthropic, Gemini-compatible APIs with Alibaba Cloud, OpenRouter and more
  • Qwen3-Coder co-evolution — both framework and Qwen3-Coder model are open-source and ship together
  • Built-in Skills and SubAgents — rich built-in tools for full agentic workflow with Claude Code-like experience
  • Terminal-first IDE-friendly — designed for CLI developers with optional VS Code, Zed, and JetBrains IDE integration
  • Multi-language docs and auth — supports Chinese, English, German, French, Japanese with API Key and Coding Plan auth
  • Centralized settings management — unified model, API key, and default behavior configuration via ~/.qwen/settings.json

Use Cases

💡 Terminal AI coding: use Qwen3-Coder in the command line for code generation and understanding
💡 Codebase navigation and refactoring: understand large codebase structure and automate refactoring operations
💡 Automated unit test generation: generate test cases for existing code modules
💡 Multi-file batch editing: make code changes across multiple files with synchronized updates
💡 Team development workflow integration: provide unified AI coding capabilities via Coding Plan subscription

Strengths & Limitations

Strengths

  • Actively maintained, recent updates
  • High community interest (27.6k stars)
  • Permissive open-source license (Apache-2.0)

Quick Start

```bash
curl -fsSL https://qwen-code-assets.oss-cn-hangzhou.aliyuncs.com/installation/install-qwen-standalone.sh | bash
qwen
/auth
```

Related Projects

Mini SWE-Agent

6.9k · Python
Active A+

Mini SWE-Agent is a minimalist AI agent in just 100 lines of code that solves GitHub issues or assists developers in the command line, demonstrating core coding agent capabilities with minimal implementation.

codingpythonagent +2
  • · Minimal ~100-line Python agent class achieving >74% on SWE-bench verified
  • · Bash-only tool interface requiring no special LLM tool-calling capabilities
  • · Completely linear message history for easy debugging and fine-tuning

Related Articles