qwen-code
An open-source AI coding agent for terminal, desktop, editors, and chat platforms using local or API-based models.
Run automated agent teams, custom skills, and multi-model code workflows across terminal, IDEs, and chat.
Qwen Code is an open-source AI coding agent that runs in your terminal, IDE, desktop app, or chat platforms. It provides agentic workflows including automated memory, sub-agents, and Model Context Protocol (MCP) support. Developers seeking a multi-model alternative to proprietary coding agents can run it across various LLM providers.
What it is
Qwen Code manages software development tasks through terminal commands, workspace edits, sub-agent teams, and LSP integration. It connects to multiple model providers like Qwen, Anthropic, OpenAI, or local instances using Ollama and vLLM. Beyond terminal commands, it operates through desktop applications, web interfaces, and messaging bots.
Who it's for
It suits software developers and automation engineers who want full control over their AI coding assistant and model configuration. It fits teams wanting self-hosted or multi-provider agent workflows across CLI and IDEs. It is not designed for non-technical users who require zero terminal setup or no API key configuration.
Key features
Multi-Provider Model Support
Switch between Qwen, OpenAI, Anthropic, Gemini, or local Ollama and vLLM instances at runtime.
Agentic Workflow Suite
Execute complex development tasks using SubAgents, Agent Teams, Auto-Memory, and built-in review skills.
Cross-Interface Operations
Run through terminal sessions, VS Code, JetBrains, Zed, web browser, or chat bots like Telegram and Feishu.
Multi-Model Agent Arena
Benchmark and compare different language models head-to-head on the exact same coding task.
Up and running in 3 steps
Our honest verdict
What's great
- Fully open-source framework supporting both cloud API providers and local models via Ollama or vLLM.
- Includes built-in sub-agents, agent teams, and computer use desktop automation capabilities.
- Integrates across terminal, IDEs, headless script execution, and chat applications like Telegram.
- Features built-in Model Context Protocol (MCP) and Language Server Protocol (LSP) integrations.
Where it falls short
- Requires paying for third-party LLM API keys or providing high-spec hardware to run local models.
- Requires Node.js 22+ or running shell installation scripts across target operating systems.
- Daemon mode and Web UI features are currently marked as experimental.
Pricing & license
Qwen Code is free and open-source under the Apache-2.0 license. Operating costs depend on the API model providers you connect to or the hardware costs associated with running local models via vLLM or Ollama.
Open remote AI-training roles
874 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I run Qwen Code with locally hosted models?
Yes, Qwen Code supports local models hosted on Ollama or vLLM alongside third-party cloud APIs.
Which IDEs and editors are supported?
Qwen Code provides integrations for VS Code, JetBrains IDEs, and Zed.
Does Qwen Code support background automation or CI/CD integration?
Yes, you can run Qwen Code in headless mode with command arguments or use the Python, TypeScript, and Java SDKs.
