How to Connect Your AI Assistant to eRS

MCP stands for Model Context Protocol, an open standard that lets AI assistants talk to external tools securely. The eRS MCP Server is eRS's implementation of that standard: a thin layer that translates a natural-language request into an eRS action and returns the result. It does not replace the eRS web application or store data separately.

The eRS MCP Server connects eResource Scheduler to AI assistants like Claude, ChatGPT, Microsoft Copilot, Cursor, and more. Once connected, an AI assistant can search resources, check availability, log timesheets, or run a utilization report directly from a chat window, without opening eRS.

Two things determine what the assistant can do: 

• The tools registered on the eRS MCP Server.
• The permissions on the authenticated eRS account. 

If an action isn't possible in eRS, it isn't possible through the assistant either.

Before You Start

Before setting up the integration, make sure the following are in place:

• An active eRS account with a role that has access to the required data
• MCP enabled for that eRS account by an eRS admin
• An MCP-compatible AI assistant or development tool (Claude Desktop, Microsoft Copilot, Cursor, ChatGPT, Google Antigravity, or any client that supports remote Streamable HTTPS and OAuth)
• The eRS MCP Server endpoint URL: https://test.eresourcescheduler.cloud/mcp

Important: eRS passwords, API keys, and access tokens should never be pasted into a client configuration file. Authentication happens through eRS OAuth in the browser, not through pasted credentials.

Connecting An AI Assistant 

Setup always comes down to two steps. Add the eRS MCP link to the assistant's settings, then sign in with an eRS account.

Claude

To set up the eRS MCP Server in Claude:

1. Open Claude and click on the + button in the chatbox. 

   

2. Click on Add Connector > Add custom connector.

3. Add the name for the connector and use the eRS MCP Server URL.

   

4. Click Continue and keep the Authentication & OAuth Client settings to default only.

5. Click Add.

   

6. Complete the eRS authentication flow when prompted.

7. Once authenticated, the eRS MCP connector is available in Claude.

   

Microsoft Copilot Studio

To set up the eRS MCP Server in Microsoft Copilot Studio:

Prerequisites: 

• A Microsoft Copilot Studio account
• Generative orchestration enabled in the agent

Note: These steps are based on the new-experience toggle being off.

1. Open Microsoft Copilot Studio and click on Create blank agent. 

   

2. Wait for the agent to finish provisioning (status shows “Ready”), then go to the Tools tab.

   

3. Click Add tool > Add new MCP. 

   

4. Enter the server name and server URL, and set Authentication to OAuth 2.0 with dynamic discovery.

   

5. Click the "No connection" dropdown and create a new connection.

   

6. Test the connection in the testing pane on the right by entering a query, for example: "List Project Apollo's details."

   

7. This prompts a new connection to be created and opens the connection manager. Once the connection is created, click “Retry”. The agent should now be able to retrieve eRS information.

8. Publish the agent to make it available across the organization, including Teams, Microsoft 365 Copilot, and other Copilot surfaces.

   

Cursor

To set up the eRS MCP Server in Cursor:

1. Open Cursor > Settings > Customize.

   

2. Click the MCPs tab > New button, then add the URL and save it. 

   {
     "mcpServers": {
       "eResource Scheduler": {
         "url": "https://test.eresourcescheduler.cloud/mcp"
       }
     }
   }

   

   Once saved, the entry looks like this:

   

3. Find the eResource Scheduler and click Authenticate.

   

4. Complete the eResource Scheduler authentication flow in the browser.

5. Confirm the eResource Scheduler shows as connected.

   

ChatGPT

To set up the eRS MCP Server in ChatGPT: 

1. Open Settings in ChatGPT.

2. Go to Plugins.

   

3. Click Add > Add MCP server.

4. Name it "eResource Scheduler" and select Streamable HTTP as the transport type.

   

5. Enter the MCP Server URL: https://test.eresourcescheduler.cloud/mcp and click Save (other fields are optional and can be left as they are).
6. Go to the MCPs tab and click Authenticate.

7. Complete the eRS authentication flow.

   

8. Confirm the eRS MCP server is active.

GitHub Copilot

To set up the eRS MCP Server in GitHub Copilot: 

1. Open Settings.

2. Select Customize, then Go to Customize.

   

3. Select Add.

   

4. Select MCP server.

5. Enter the server name and select HTTP as the type.

   

6. Add the following URL: https://test.eresourcescheduler.cloud/mcp and click on ‘Add server’.

7. Go to the MCP tab in Customize, find it under Installed, and complete the authentication process.

   

8. Confirm the authentication process is complete.

   

Google Antigravity

To set up the eRS MCP Server in Google Antigravity: 

1. Open Settings and go to Customizations.

   

2. Select MCP Server, then open the MCP configuration file.
3. Add this configuration below:

{
  "mcpServers": {
    "ers": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://test.eresourcescheduler.cloud/mcp"
      ]
    }
  }
}

4. Save the configuration and authenticate with eRS when prompted.

5. Confirm the eRS MCP Server is authenticated. The eRS tools should now be visible.

   

Other MCP-Compatible Clients

Most MCP-compatible assistants and dev tools can connect to eRS using standard settings:

• Transport: Streamable HTTP
• Server URL: https://test.eresourcescheduler.cloud/mcp
• Authentication: OAuth

If the client requires mcp-remote, use:

• Command: npx
• Arguments: -y mcp-remote https://test.eresourcescheduler.cloud/mcp
• Environment: None

For the full list of available tools and prompting best practices, see eRS MCP Server: Tools & Prompting Guide.

Troubleshooting

Note: eRS does not control how a connected AI client stores or uses retrieved data. Check that tool's own data retention and training settings if this matters to the organization.