Use Qwen Code with any AI provider

Alibaba Qwen Code CLI — supports OpenAI, Anthropic & Gemini providers via kRouter. Pointing Qwen Code at kRouter gives it 100+ providers behind a single local endpoint — free tiers, subscriptions you already pay for, and paid APIs — with automatic failover when one runs out. Qwen Code is configured through its own settings UI.

Setup

1. 1

   Install Qwen Code

   npm install -g @qwen-code/qwen-code

2. 2

   API Key

3. 3

   Base URL

   {{baseUrl}}
4. 4

   Select Model

5. 5

   Save Config

   Copy the JSON below to your ~/.qwen/settings.json file.

{
  "security": {
    "auth": {
      "selectedType": "openai",
      "apiKey": "{{apiKey}}",
      "baseUrl": "{{baseUrl}}"
    }
  },
  "model": {
    "name": "{{model}}"
  }
}

Qwen Code documentation

What changes for Qwen Code

Nothing about how you use Qwen Code changes — same interface, same workflow. What changes is where its requests land. Instead of being tied to one vendor and one bill, every request goes through your own machine first and out to whichever provider you mapped: a free tier for routine edits, a subscription you already pay for on the heavy turns, a paid API only when nothing else fits. When one account hits a limit, the next takes over mid-session rather than failing the task. And because the router runs locally, your prompts are not passing through a third-party service on the way.

Troubleshooting Qwen Code

Qwen Code + kRouter FAQ