Register an EcoToken model in models.json so WorkBuddy/CodeBuddy can use the OpenAI-compatible API.

01 Prepare access details

Create an API key that can access deepseek-v4.1-flash. Supply the real key through ECOTOKEN_API_KEY.

02 Install the client

Install the desktop app or IDE extension from the WorkBuddy/CodeBuddy website. Skip this step if it is already installed.

03 Configure the model

Create user-level ~/.codebuddy/models.json or project-level .codebuddy/models.json. Project configuration takes priority, and the file must be UTF-8.

{
  "models": [
    {
      "id": "deepseek-v4.1-flash",
      "name": "deepseek-v4.1-flash",
      "vendor": "EcoToken",
      "url": "https://ecotoken.ecophase-ai.com/v1/chat/completions",
      "apiKey": "${ECOTOKEN_API_KEY}",
      "maxInputTokens": 128000,
      "maxOutputTokens": 8192,
      "supportsToolCall": true,
      "supportsImages": false
    }
  ],
  "availableModels": ["deepseek-v4.1-flash"]
}

Set context, output, image, and tool-call capabilities to the actual model capabilities shown in the EcoToken console.

04 Configure the API key

export ECOTOKEN_API_KEY="YOUR_API_KEY"

Restart WorkBuddy/CodeBuddy after using setx on Windows.

05 Launch and verify

Restart the client, select deepseek-v4.1-flash, and send a short prompt. A model response and matching entry under Usage & costs → Request records confirm the integration.

06 Common errors

• 401: Check that ECOTOKEN_API_KEY is available to the client process.
• 403: Check the key group, model permission, and account state.
• Base URL: The url field requires the full https://ecotoken.ecophase-ai.com/v1/chat/completions request URL.