Use TeamoRouter in VS Code GitHub Copilot Chat: Custom Endpoint (Bring Your Own Key)

GitHub Copilot Chat in VS Code supports Bring Your Own Key (BYOK): add a Custom Endpoint with your TeamoRouter API key and endpoint URL, and Claude, GPT and other models show up in chat and agent mode. Custom Endpoint supports three API types — Chat Completions, Responses and Anthropic Messages — and TeamoRouter serves all three.

1. Before you start

• Update VS Code to a current version. Custom Endpoint replaces the old "OpenAI Compatible" provider and the github.copilot.chat.customOAIModels setting.
• Create an API key in the TeamoRouter dashboard.
• In Local sessions, BYOK models work without signing in to GitHub and without a Copilot plan. On Copilot Business or Enterprise, your admin may have turned off BYOK on GitHub, so check first.

2. Add a Custom Endpoint

1. In the Chat view, open the model picker and select the gear (Manage Language Models), or run Chat: Manage Language Models from the Command Palette.
2. Select Add Models, then Custom Endpoint.
3. Enter TeamoRouter as the group name (the label the model picker groups by).
4. Enter any display name, and paste your TeamoRouter key as the API key.
5. Pick Messages as the API type (for Claude models). You can override it per model, as shown below.
6. VS Code opens chatLanguageModels.json. Keep the apiKey line it generated, change models to the following, and save:

[
  {
    "name": "TeamoRouter",
    "vendor": "customendpoint",
    "apiKey": "(keep the line VS Code generated)",
    "apiType": "messages",
    "models": [
      {
        "id": "claude-sonnet-5-5",
        "name": "Claude Sonnet 5.5 · TeamoRouter",
        "url": "https://api.teamorouter.com/v1/messages",
        "apiType": "messages",
        "toolCalling": true,
        "vision": true,
        "contextWindow": 1000000,
        "maxOutputTokens": 32000
      },
      {
        "id": "claude-opus-5-5",
        "name": "Claude Opus 5.5 · TeamoRouter",
        "url": "https://api.teamorouter.com/v1/messages",
        "apiType": "messages",
        "toolCalling": true,
        "vision": true,
        "contextWindow": 1000000,
        "maxOutputTokens": 32000
      },
      {
        "id": "gpt-6.1-sol",
        "name": "GPT-6.1 Sol · TeamoRouter",
        "url": "https://api.teamorouter.com/v1/responses",
        "apiType": "responses",
        "statefulResponses": false,
        "toolCalling": true,
        "vision": true,
        "contextWindow": 1000000,
        "maxOutputTokens": 32000
      }
    ]
  }
]

7. Back in chat, pick a model from the TeamoRouter group in the model picker. If it doesn't appear, restart VS Code.

3. Which API type to use

• Give the full endpoint URL (ending in /messages, /responses or /chat/completions); VS Code uses it as is.
• Add "statefulResponses": false for the Responses API. By default VS Code sends only new messages and relies on previous_response_id; turning it off sends the full conversation each time and avoids errors about missing tool calls.
• To add another model, copy a block and change id and name. Model IDs are in the model catalog. Agent mode lists only models with toolCalling set to true.

Troubleshooting

• Inline completions (Tab) still need GitHub: BYOK covers chat and small background jobs such as title generation. Inline suggestions, semantic search and similar features still need a GitHub account.
• The model doesn't show in the picker: make sure chatLanguageModels.json is saved and is valid JSON, then restart VS Code.
• Requests return 401: the key is wrong or was deleted. Edit the group in Language Models and paste the key again.
• You use OpenAI's Codex extension, not Copilot: the Codex extension reads ~/.codex/config.toml. Set it up with the Codex CLI guide.

Related

• Connect VSCode + Cline
• Connect Codex CLI
• API integration