Skip to main content

Connect Lexeris MCP server to GitHub Copilot

Stefan Rohde
  • Updated

This guide describes how to set up the Lexeri MCP Server (Model Context Protocol) manually in GitHub Copilot. Once configured, GitHub Copilot Chat can access your Lexeri termbase directly – for example to search for terms, check texts for terminology violations, or create new term suggestions.

Note: MCP server support in GitHub Copilot is currently in public preview and is available in Visual Studio Code (VS Code) as well as other supported IDEs.
 

Prerequisites

  • Visual Studio Code (version 1.99 or later) with the GitHub Copilot extension installed
  • An active GitHub Copilot subscription
  • An existing Lexeri account with at least one termbase
  • A valid Lexeri API token (found in your Lexeri termbase settings)
  • Organisation administrator rights in GitHub (required once to enable MCP for the entire organisation)

Step 1: Enable MCP servers in the GitHub organisation (administrator)

This step must be completed once by a GitHub organisation administrator. Without this setting, members of the organisation cannot use MCP servers in GitHub Copilot – even if the local configuration is correct.

  1. Open the Copilot policies of your GitHub organisation at https://github.com/organizations/YOUR-ORGANISATION/settings/copilot/policies.
  2. In the Policies section, find the setting MCP servers in Copilot.
  3. Set the value to Enabled and save the change.
Note: This setting applies to all members of the organisation. Once enabled, all users with an active Copilot subscription can set up MCP servers in their local VS Code configuration.
 
 

Step 2: Open or create the MCP configuration file

GitHub Copilot reads MCP server configurations from a central file. Open the command palette with Ctrl+Shift+P (Windows/Linux) or Cmd+Shift+P (macOS) and enter the following command:

MCP: Open User Configuration File

VS Code will open the mcp.json file in your user directory. If the file does not yet exist, it will be created automatically.

Note: Alternatively, you can create a project-specific configuration by adding the file .vscode/mcp.json to your project folder. This configuration will then apply only to that project.
 
 

Step 3: Configure the Lexeri MCP server

Add the following entry to your mcp.json file.

{
  "servers": {
    "lexeri": {
      "type": "http",
      "url": "https://mcp.lexeri.com/mcp"
    }
  }
}

Then save the file with Ctrl+S (Windows/Linux) or Cmd+S (macOS).

Step 4: Connect to the MCP server

After saving the configuration file, a "Start" link will appear above the "lexeri" entry in the file. Click it to establish the connection to the MCP server.

Step 5: Start the MCP server in VS Code

After saving the configuration, open the command palette again (Ctrl+Shift+P / Cmd+Shift+P) and run the following command:

MCP: List Servers

lexeri should appear in the list. Click on it and select Start Server to establish the connection. A green icon indicates that the server has connected successfully.

Step 6: Use Lexeri tools in GitHub Copilot Chat

Open GitHub Copilot Chat in VS Code via the Chat icon in the sidebar or with Ctrl+Alt+I / Cmd+Alt+I. Make sure the Agent mode is selected (not "Ask" or "Edit").

Click the Tools icon (wrench icon) in the input field to verify that the Lexeri tools are available. You can then submit requests such as:

  • "Search my Lexeri termbase for the term 'Renewable Energy'."
  • "Check the following text for terminology violations according to my Lexeri termbase."
  • "Create a term suggestion for the term 'Carbon Footprint' in Lexeri."

Related articles