Qwen Code and CodeWhale Integration Guide

Applicable roles: Developers, Admin Users, Member Users Last updated: 2026-08-06

These two tools have different configuration methods. The current version of Qwen Code supports adding a custom Provider in modelProviders. CodeWhale's field structure may change across releases — check the help documentation and example files provided by your local version before configuring.

Qwen Code

Install and set the key:

npm install -g @qwen-code/qwen-code@latest
export CUSTOM_API_KEY="YOUR_API_KEY"

Register the platform as an OpenAI Compatible Provider in ~/.qwen/settings.json:

{
  "modelProviders": {
    "myplatform": [
      {
        "id": "deepseek-v4-flash",
        "name": "DeepSeek V4 Flash",
        "envKey": "CUSTOM_API_KEY",
        "baseUrl": "https://<your-api-endpoint>"
      }
    ]
  },
  "providerProtocol": {
    "myplatform": "openai"
  }
}

The current specification requires the Provider's value to be a model array. Older tutorials that define the Provider as an object with a models field may be silently ignored in newer versions. To add more models, append entries at the same level within the myplatform array.

Use the API endpoint shown on the console "API Key" page.

After launching, use /model to select the corresponding model. The API key should not be written directly into the settings file — the environment variable referenced by envKey is read at runtime.

CodeWhale

First, check whether your current version of CodeWhale offers an OpenAI Compatible or Custom Provider option. If you can fill in the following three fields, use:

Base URL: https://<your-api-endpoint>
API Key: YOUR_API_KEY
Model: deepseek-v4-flash

If the official examples explicitly use TOML, fill in the same values using the corresponding field names. Do not directly reuse configuration paths or environment variable names from other versions, as the tool may fail to read them without giving an obvious error.

Verification Steps

First, confirm the call name in the model selector, then send a short message and check the console usage records. If the model does not appear, the issue is usually with the local configuration structure or file location. If the model is selectable but requests fail, check the key, Base URL, and group permissions.