Integrate WorkBuddy

WorkBuddy/CodeBuddy can load custom models from a local models.json file. Point each model to the HiModels OpenAI Chat Completions-compatible endpoint to use models enabled for your account.

1. Prepare WorkBuddy/CodeBuddy

  1. Install and sign in to WorkBuddy/CodeBuddy.
  2. Open a project directory at least once so the app creates the local .codebuddy configuration directory.
  3. Create an API Key from HiModels API Keys.

2. Configure Local Models

Create or edit the user-level configuration file:

C:\Users\<your-user-name>\.codebuddy\models.json

To limit the configuration to one project, use the project-level file instead:

<your-project>\.codebuddy\models.json

Set the HiModels API Key environment variable:

# Windows
setx HIMODELS_API_KEY "<your_himodels_api_key>"

# macOS / Linux
export HIMODELS_API_KEY="<your_himodels_api_key>"

Then configure models.json with HiModels model IDs and endpoint:

{
  "models": [
    {
      "id": "gpt-5.5",
      "name": "GPT-5.5 (HiModels)",
      "vendor": "HiModels",
      "url": "https://api.himodels.ai/v1/chat/completions",
      "apiKey": "${HIMODELS_API_KEY}",
      "maxInputTokens": 128000,
      "maxOutputTokens": 8192,
      "supportsToolCall": true,
      "supportsImages": true
    },
    {
      "id": "deepseek-v4-pro",
      "name": "DeepSeek V4 Pro (HiModels)",
      "vendor": "HiModels",
      "url": "https://api.himodels.ai/v1/chat/completions",
      "apiKey": "${HIMODELS_API_KEY}",
      "maxInputTokens": 128000,
      "maxOutputTokens": 8192,
      "supportsToolCall": true,
      "supportsImages": false
    }
  ],
  "availableModels": [
    "gpt-5.5",
    "deepseek-v4-pro"
  ]
}

Save models.json as UTF-8 without BOM. Adjust model IDs and token limits according to HiModels Explore and your SKU configuration.

3. Restart and Select a Model

  1. Completely quit WorkBuddy/CodeBuddy and reopen it.
  2. Choose a custom model whose name includes “HiModels” from the model selector.

4. Optional: Verify the API Key

Call the HiModels endpoint directly first to verify the API Key, model ID, and network:

curl https://api.himodels.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${HIMODELS_API_KEY}" \
  -d '{"model":"gpt-5.5","messages":[{"role":"user","content":"hi"}],"stream":false}'

For 401, check the API Key. For 404, check the model ID and endpoint. If models do not appear, validate the JSON and file path, then fully restart the app.