Overview

Most marketers spend a fair amount of time clicking through reports, exporting CSVs, and stitching numbers together to answer simple questions. The MCP server skips that step.

• Ask in plain English. No filters to set, no date pickers, no API documentation to read.
• Get answers in seconds. "What was the CTR on last week's welcome email?" returns a number, not a five-step navigation path.
• Cross-campaign comparisons get easier. The assistant can pull stats for several campaigns at once and put them side by side.
• Diagnosing problems is faster. Instead of opening the funnel report manually, you can ask why delivery dropped and get a useful starting point.
• The creative is one prompt away. Pull up the HTML, push payload, or template body without hunting for the campaign in the dashboard.

It's a layer on top of the same Netcore CE data you already have.

Environment URL

Access from Netcore Panel

Before any AI client (ChatGPT, Cursor, or Claude) can connect, MCP must be enabled for the workspace from the Netcore panel.

1. Log in to the Netcore CE panel and navigate to Account configuration.
2. Under MLAI Features and Enable MCP
3. Review and select Accept to grant access, or Cancel to back out without enabling MCP.

Understand MCP Flow

MCP is an open standard that lets AI assistants connect to external tools and services. Instead of opening the dashboard or writing API calls, you describe the outcome in plain English. Your AI assistant picks the right tool, and the MCP server handles authentication, the API request, and the response.

You (plain English) → AI Assistant → MCP Protocol → Netcore CE MCP Server → Netcore CE API

Channels Supported

Available Tools

The Netcore CE MCP server includes 15 read-only tools that help you search, view, and analyse campaign-related data without making any changes to your account. This is the starting point for most questions, find the campaigns first, then pull their stats.

Channel-Specific Tools

Use these tools when you want to go deeper into a single campaign, review its full setup or check its performance for that channel.

Authentication

The Netcore CE MCP server uses OAuth-based authentication tied to your Netcore CE account.

When you connect for the first time, you are redirected to authenticate with your Netcore CE credentials. If you are already logged into Netcore CE, you see an authorisation prompt requesting permission to access your workspace data.

The authorisation screen displays:

• Account: Netcore CE email.
• Workspace: The workspace you are currently logged into
• Role: It describes the role in the current workspace, for example, Admin, Manager, or Analyst
• Data Centre: Your data centre environment
• Permissions requested: The scopes the MCP server requires, for example Campaigns:read

Click Accept to grant access.

Key Authentication Behaviours

• The MCP server inherits the environment, workspace, and role of your current Netcore CE session.
• Authentication tokens are derived from your active Netcore CE session.
• If you log out or switch workspaces in Netcore CE, you are prompted to re-authenticate the MCP connection.
• All data access respects your existing Netcore CE role-based permissions (RBAC).

UI Setup

Connect the Netcore CE MCP server to your preferred AI assistant. All three clients use the same OAuth sign-in, you authenticate once with your Netcore CE credentials, and the assistant takes it from there.

MCP Server URL

https://cee-mcp.netcorecloud.com/mcp

Connect with Claude

Pro / Max users

1. Open Claude and navigate to Settings > Connectors.
2. Click ＋ Add custom connector.
3. Select Web Remote MCP as the connector type.
4. Enter the MCP server URL from here.
5. Click Add to save the connector.
6. Click Connect and complete sign-in on the Netcore login page in your browser.
7. Return to Claude and confirm the connector is enabled.
8. You can then set Tool permissions for each tool: Automatic, Ask first, or Disabled.

   📘

   Note

   On Team and Enterprise plans, individual users cannot add custom connectors themselves — the connector must first be added at the organization level.

Team Enterprise Users/ Enterprise Users

1. Navigate to Organization Settings > Connectors.
2. Adds the custom connector URL for the workspace.
3. Each user has to navigate to Settings > Connectors and clicks Connect for the Netcore CE connector to complete their individual sign-in.

Once connected, you can start asking Claude questions like:

• "Show me all running email campaigns."
• "What is the delivery rate for my last app push campaign?"
• "Compare click performance across my SMS campaigns from last week."

Connect with ChatGPT

1. Go to chatgpt.com and sign in.

2. Open Settings and navigate to Apps & Connectors. Depending on your rollout, this may appear as Connectors.

3. If your workspace requires it, enable Developer Mode under Advanced Settings and click Create (or Add app / Add connector). Enter the connector details:

   Name: Netcore CE MCP (or any label you prefer)

   Connector URL: the MCP server URL from the Environment URLs table above

4. Click Save to create the connector.

5. Click Connect and complete sign-in on the Netcore login page in your browser.

6. Return to ChatGPT and confirm the app shows as Connected with the Netcore CE tools visible.

Once connected, you can start asking ChatGPT questions like:

• "Search for all completed email campaigns with 'Welcome' in the name."
• "Get detailed stats for campaign ID abc123 for the last 7 days."
• "Show me which devices performed best for my latest app push campaign."

Connect with Cursor

1. Open Cursor and go to Cursor Settings.
2. Navigate to Tools & MCP.
3. Click Add new MCP server.
4. Choose Streamable HTTP (remote server) as the transport type.
5. Enter the MCP server URL from the Environment URLs table above.
6. Click Install / Save.
7. When Cursor prompts for authentication, click Connect.
8. Complete sign-in on the Netcore login page in your browser and return to Cursor.
9. Verify from the MCP settings that the server is healthy and the tools are listed.

Configuration Example

For MCP-compatible AI agents that use a JSON configuration file (such as Claude Desktop or Cursor), use the snippet below.

{
  "mcpServers": {
    "netcore-ce": {
      "type": "streamableHttp",
      "url": "https://mcp.netcorecloud.com"
    }
  }
}

Authentication is handled via the OAuth redirect described above. Credentials are never passed as headers in your local configuration.

Transport Options

Example Queries

Here are some examples of what you can ask your AI assistant once connected.

Agent Permission and Access Control

When managing AI agent permissions, apply stricter controls for any task that affects production data or sends outbound communication. Although the Netcore CE MCP server is read-only today, the principle still applies as new tools are added over time.

The risks of granting permanent, unrestricted permissions to an AI agent include:

• Exposing campaign-level data without human oversight during unexpected behaviour or prompt injection attempts.
• Bypassing approval workflows that protect compliance and brand standards.
• Generating noisy, irrelevant queries that consume API quota.

Refer to the MCP implementation guidelines for more details.

Built-in Safety Controls

The Netcore CE MCP server includes several controls to reduce operational risk:

• Read-only scope: The server cannot create, modify, or delete campaigns, segments, or templates.
• Session-scoped credentials: Credentials are derived from your active Netcore CE session and are never exposed in model-visible content.
• Input validation: Runs before every upstream API call. Missing or malformed parameters return a guided error rather than a silent failure.
• HTTPS and timeouts: Protect all upstream communication.
• Rate limiting: Configurable per workspace.

Security and Permissions

• The MCP server provides read-only access to your campaign data. It cannot create, modify, or delete campaigns.
• All data access is scoped to the workspace and role of the authenticated user.
• Data shared with AI assistants is subject to the respective AI provider's data-handling policies.
• Review Netcore's Privacy Policy and Data Security pages for more information.

Limitations

• Date range: Analytics queries are limited to a maximum of 30 days per request.
• Batch size: get_campaign_stats supports up to 50 campaign IDs per request.
• Content size: Email HTML content can be large (10–50 KB). Use get_campaign_content only when you need to analyse the actual creative.
• Channel availability: The MCP server only returns data for channels enabled on your workspace. Contact the Netcore helpdesk to enable additional channels.

Warning

Troubleshooting & FAQs

Q. I cannot access the MCP server?

A. If you cannot access the MCP server or your campaign data, check your workspace connection. The MCP server inherits the active session from your Netcore CE account.

To resolve this issue:

1. Verify that you are signed in to the correct Netcore CE workspace.
2. If you have signed out or switched to a different workspace, re-authenticate the MCP server connection to restore access to your campaign data.

Q. My tool not appearing in the AI assistant.

A. If a specific tool is missing after connecting:

1. Restart your AI assistant session.
2. Confirm that the channel associated with the tool (for example, WhatsApp or RCS) is activated on your workspace.
3. Confirm that your role has read access to the relevant module.

Empty or stale results

If a query returns no data or appears stale:

1. Ensure the date range is within the last 30 days.
2. Confirm that the campaign exists in the workspace you are currently authenticated against.
3. Re-run the query after a few minutes — analytics may take time to aggregate for very recent campaigns.