GitHub Copilot

Use Cortecs models in GitHub Copilot Chat

Visual Studio Code supports custom model endpoints for GitHub Copilot Chat. Connect Cortecs to use its models for chat and agent tools in VS Code.

Before you begin, create a Cortecs API key by following the Quickstart and choose a tool-calling model from the model catalog.

1. Add a custom endpoint

1. Install the latest Visual Studio Code and the GitHub Copilot Chat extension.

2. Open the Command Palette and run Chat: Manage Language Models.

3. Select Add Models, then Custom Endpoint.

4. Use Cortecs as the group and display name, enter your Cortecs API key, and select Responses as the API type.

2. Configure a model

VS Code opens chatLanguageModels.json. Update the generated model entry with:

• id: The exact Cortecs model ID

• name: The label shown in the model picker

• url: https://api.cortecs.ai/v1/responses

• toolCalling: true for a model that supports tool calling

• vision: true only for a model that supports image input

• maxOutputTokens: A supported output limit for the model

• maxInputTokens: The model's context size minus maxOutputTokens

Use the model catalog to check capabilities and context size. Keep the API key value generated by VS Code instead of adding the key directly to version-controlled files.

Save the file, return to Copilot Chat, and select the Cortecs model from the model picker. If it does not appear, reload VS Code.

The Cortecs Responses API is recommended for reasoning, tool-calling, and multi-turn agent workflows. VS Code also supports Chat Completions; select it and use https://api.cortecs.ai/v1/chat/completions only if a model or workflow requires that API.

Custom endpoints power Copilot Chat and agent features, but not inline code completions, semantic search, or other embedding-based features.

For the full configuration schema, see the Visual Studio Code language model documentation.