VS Code

Use VS Code's BYOK Custom Endpoint provider to add Voxvey models to Copilot Chat and agent mode.

1. Open the model manager

1. Open the Command Palette.
2. Run Chat: Manage Language Models.
3. Select Add Models.
4. Select Custom Endpoint.
5. Enter Voxvey as the group name.
6. Enter your Voxvey token when prompted for the API key.

The Custom Endpoint provider supports Chat Completions, Responses, and Messages API types. Voxvey exposes each of those API surfaces directly.

2. Configure models

VS Code opens chatLanguageModels.json. Add these Voxvey model entries.

[
  {
    "name": "Voxvey",
    "vendor": "customendpoint",
    "apiKey": "YOUR_VOXVEY_TOKEN",
    "models": [
      {
        "id": "deepseek/deepseek-v4-flash",
        "name": "Voxvey DeepSeek V4 Flash",
        "url": "https://api.voxvey.com/v1/chat/completions",
        "apiType": "chat-completions",
        "toolCalling": true,
        "vision": false,
        "maxInputTokens": 1000000,
        "maxOutputTokens": 384000,
        "streaming": true,
        "supportsReasoningEffort": ["low", "medium", "high"]
      },
      {
        "id": "<messages-model-id>",
        "name": "Voxvey Messages API",
        "url": "https://api.voxvey.com/v1/messages",
        "apiType": "messages",
        "toolCalling": true,
        "vision": false,
        "maxInputTokens": 200000,
        "maxOutputTokens": 64000,
        "streaming": true,
        "thinking": true
      },
      {
        "id": "openai/gpt-5.5",
        "name": "Voxvey GPT-5.5",
        "url": "https://api.voxvey.com/v1/responses",
        "apiType": "responses",
        "toolCalling": true,
        "vision": false,
        "maxInputTokens": 400000,
        "maxOutputTokens": 64000,
        "streaming": true,
        "thinking": true,
        "supportsReasoningEffort": ["low", "medium", "high"]
      }
    ]
  }
]

3. Select the model

Open Chat, use the model picker, and choose either:

Voxvey DeepSeek V4 Flash
Voxvey Messages API
Voxvey GPT-5.5

If the models do not appear, restart VS Code.

Notes

• Models must support tool calling to appear in agent mode.
• Use the Messages API entry for clients and workflows that expect /v1/messages.
• Use the Responses model for richer agent workflows.
• Use the Chat Completions model for fast compatibility testing.