Integrate with Qwen Code

Qwen Code is an open-source AI agent that lives in your terminal, built by the Qwen team at Alibaba Group. It has DeepSeek as a built-in third-party provider — just bring your API key.

1. Install Qwen Code

  • Install Node.js 20 or later.
  • Install Qwen Code:
npm install -g @qwen-code/qwen-code@latest

Or via the quick-install script:

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

Or Homebrew:

brew install qwen-code
  • Verify:
qwen --version

2. Get a DeepSeek API Key

Go to the DeepSeek Platform, create an API key, and copy it.

3. Configure DeepSeek via /auth

Launch Qwen Code, then run the /auth command:

qwen

Inside the interactive session:

/auth

The /auth menu will appear. Follow these steps:

  1. Select Third-party Providers
  2. Select DeepSeek API Key
  3. Step 1/2 — paste your DeepSeek API key and press Enter
  4. Step 2/2 — confirm the model IDs: deepseek-v4-pro, deepseek-v4-flash (or edit if needed), press Enter

After confirmation, you'll see: Successfully configured DeepSeek API Key.

4. Switch to a DeepSeek Model

Run the /model command and select deepseek-v4-pro or deepseek-v4-flash:

/model

You're ready to go. Qwen Code will now use DeepSeek V4 models for coding.

Tip: Use /model to switch between models at any time. Use /auth again if you need to reconfigure your API key.